{ "source_file": "0930", "source_platform": "Medium", "author": "Lak Lakshmanan", "title_listed": "What it means to get your data ready for AI", "keywords": "数据准备,AI", "domain": "数据处理", "content_type": "深度技术解析", "is_paid_flag": "部分免费/会员墙", "url": "https://aiadvances.org/what-it-means-to-get-your-data-ready-for-ai-518861a8f025", "publish_date": "2025-11", "corpus_stage": "预训练/SFT", "notes": "", "id": "4036ef3c3f6d", "collected_at": "2026-10-02T00:39:36", "title": "Medium", "body_markdown": "[Sitemap](https://aiadvances.org/sitemap/sitemap.xml)\n[Open in app](https://play.google.com/store/apps/details?id=com.medium.reader&referrer=utm_source%3DmobileNavBar&source=---top_nav_layout_nav-------------------------------------------)\nSign up\n[Sign in](https://medium.com/m/signin?operation=login&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=post_page---top_nav_layout_nav-----------------------global_nav--------------------)\n[Medium Logo](https://medium.com/?source=---top_nav_layout_nav-------------------------------------------)\nGet app\n[Write](https://medium.com/m/signin?operation=register&redirect=https%3A%2F%2Fmedium.com%2Fnew-story&source=---top_nav_layout_nav-----------------------new_post_topnav--------------------)\n[Search](https://medium.com/search?source=---top_nav_layout_nav-------------------------------------------)\nSign up\n[Sign in](https://medium.com/m/signin?operation=login&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=post_page---top_nav_layout_nav-----------------------global_nav--------------------)\n![Unknown user](https://miro.medium.com/v2/resize:fill:64:64/1*dmbNkD5D-u45r44go_cf0g.png)\n## [AI Advances](https://aiadvances.org/?source=post_page---publication_nav-3fe99b2acc4-518861a8f025-----------------------------------------)\n·\n 1. [What’s causing this disruption?](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#7832 \"What’s causing this disruption?\")\n 2. [1. Rethink ETL/ELT: From Normalization to Context](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#8221 \"1. Rethink ETL/ELT: From Normalization to Context\")\n 3. [2. Prioritize Data Curation Over Data Collection](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#b1f0 \"2. Prioritize Data Curation Over Data Collection\")\n 4. [3. Build Agent-Ready Infrastructure: Perception and Action](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#13e0 \"3. Build Agent-Ready Infrastructure: Perception and Action\")\n 5. [4. Manage Agent Artifacts as First-Class Data](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#0a93 \"4. Manage Agent Artifacts as First-Class Data\")\n 6. [5. Connect Observation to Training](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#113c \"5. Connect Observation to Training\")\n 7. [Summary: How the role of a data engineer changes](https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------#8aa9 \"Summary: How the role of a data engineer changes\")\n\n\n[![AI Advances](https://miro.medium.com/v2/resize:fill:76:76/1*R8zEd59FDf0l8Re94ImV0Q.png)](https://aiadvances.org/?source=post_page---post_publication_sidebar-3fe99b2acc4-518861a8f025-----------------------------------------)\nIdeas in depth. Where exceptional thinking takes shape.\n[Agentic Ai](https://medium.com/tag/agentic-ai?source=post_page---header_tags--518861a8f025-----------------------------------------)\n[Data Engineering](https://medium.com/tag/data-engineering?source=post_page---header_tags--518861a8f025-----------------------------------------)\n# What it means to get your data ready for AI\n## (or) How AI Agents are changing the job of a Data Engineer\n[![Lak Lakshmanan](https://miro.medium.com/v2/resize:fill:64:64/2*TveVoapl-TEk-jBTrbis8w.jpeg)](https://lakshmanok.medium.com/?source=post_page---byline--518861a8f025-----------------------------------------)\n[Lak Lakshmanan](https://lakshmanok.medium.com/?source=post_page---byline--518861a8f025-----------------------------------------)\n7 min read\n·\nNov 6, 2025\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fvote%2Fai-advances%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&user=Lak+Lakshmanan&userId=247b0630b5d6&source=---header_actions--518861a8f025---------------------clap_footer--------------------)\n--\n12\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Frepost%2Fp%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&user=Lak+Lakshmanan&userId=247b0630b5d6&source=---header_actions--518861a8f025---------------------repost_header--------------------)\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fbookmark%2Fp%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---header_actions--518861a8f025---------------------bookmark_footer--------------------)\n[Listen](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2Fplans%3Fdimension%3Dpost_audio_button%26postId%3D518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---header_actions--518861a8f025---------------------post_audio_button--------------------)\nShare\nI’m giving the keynote at the “[Data Engineering in the Age of AI” conference](https://www.oreilly.com/live-events/data-superstream-data-engineering-in-the-age-of-ai/0642572221676/0642572221669/) on Nov. 13, 2025. My keynote will be about why Agentic AI is disrupting data engineering and how the role of a data engineer is changing because of this. This article is an expanded version of the talk, with more explanation on the shifts in technology capabilities that are causing this change in role to happen.\n### What’s causing this disruption?\nOne reason for the shift is that the way people use data is changing.\nEnterprise applications and dashboards were built by software engineers and data scientists to handle the needs of non-technical users. In turn, business analysts and end-users passively consumed that content. There might be some interaction built into the application, but these interactions involved rigid, predetermined workflows. As data engineers, the job was to supply data in formats that such applications could consume.\nPress enter or click to view image in full size\nTransition from a “Builder”-centric model, where technical users create applications, to a “Viber”-centric model, where non-technical users interact directly with data through AI agents. Graphic generated using Gemini by author.\nIncreasingly, non-technical users are interacting directly with data. They are able to vibe-code applications and tools that do exactly what they want. Existing SaaS applications are going beyond incorporating a side-by-side chat interface to more natively embedding natural language interactions using frameworks like [CopilotKit](https://github.com/CopilotKit/CopilotKit). Instead of simply using these mechanisms to repeat the rigid workflow, forward-leaning developers are embedding AI agents into their applications where each of the backend APIs are available to the agent as tool calls.\nSecondly, the center of gravity is shifting. It used to be that data was large, and so you moved the compute to the data to avoid having to move large amounts of data around. However, frontier AI models (LLMs) are now the center of gravity and AI applications are built around them.\nPress enter or click to view image in full size\nThe center of gravity has shifted, and so the technology stack has flipped\nInstead of custom-built compute going to the data, agentic AI applications use LLMs as the reasoning engine that is able to understand user intent, reason about the task, and invoke tools to carry out actions. This new wave of applications is about converting user intent directly into actions.\nHow do these two dynamics affect the job of a data engineer? Here are five principles to keep in mind as you try to get data ready for AI.\n### 1. Rethink ETL/ELT: From Normalization to Context\nData engineers today spend enormous effort normalizing data, creating clean schemas, and building transformation pipelines. The goal is to make data understandable to downstream analysts and applications.\nThis doesn’t mean ETL/ELT becomes irrelevant — making data available remains important. But you can rely on agents to interpret schemas, understand relationships, and work with data in various formats without extensive preprocessing.\nHowever, taking existing tables and slapping on a data catalog and MCP servers on top is grossly underutilizing what agentic technology is capable of. Moreover, you make the job of the AI agent much harder than it needs to be. How?\nPress enter or click to view image in full size\nFrom normalization to context, a change in data engineering paradigm. Diagram by author.\nAI agents can understand data in context. They don’t need everything pre-normalized into rigid schemas. In fact, as the number of tables grows, today’s agents struggle to correctly interpret and write correct SQL to join all that data. Moreover, as you increase the number of narrow slices of the data, the odds of collision and ambiguity increase. For example, you may have a “loan amount” column in two tables. In one table, it may be the amount that is being requested by the borrower, and in the other table it may be the principal that was actually disbursed by the lender. The context is hard to carry over the more massaged and normalized and widespread the data landscape becomes.\nMaintain your data availability workflows, but question whether every normalization step is still necessary. Can an agent understand this data with proper context instead of transformation? Instead of the principal being a single number, can it be a paragraph from the original termsheet or funding memo that explains the tranches in which this principal will be released?\nAvoid the temptation to open up only your unstructured data to AI agents — it is easy to throw a RAG over your PDFs, emails, etc. but the really actionable data in your organization is often still the structured kind.\n### 2. Prioritize Data Curation Over Data Collection\nIn-context learning makes curation more important than collection.\nIn the big data era, the goal was collecting as much data as possible because you wanted to train machine learning models on extremely large datasets — more data meant better machine learning models.\nHowever, AI agents are built using in-context learning where you provide an example or two in the prompt. The LLM can effectively mimic these examples, whether it is to follow a process ([Chain of Thought](https://arxiv.org/abs/2201.11903)) or to follow a format or style ([Few-shot Learning](https://www.promptingguide.ai/techniques/fewshot)). With the advent of in-context learning, quality exemplars matter more than volume.\nPress enter or click to view image in full size\nThe new technology stack that you should become familar with\nThe data you show an agent as an example shapes how it understands all similar data. You might find yourself creating an [Example Store](https://docs.cloud.google.com/agent-builder/agent-engine/example-store/overview), and choosing which examples to use for a specific type of user query. With the increasing importance of data curation, it becomes important as data engineers that you build tooling to:\n * Identify your highest-quality data, examples that are complete, accurate, and representative\n * Update these exemplars regularly as standards evolve\n * Validate that your curated data actually serves as effective examples for agent learning\n\n\nA key persona that you, as a data engineer need to enable, is a data curator. The types of storage you will support also change, to include graph and vector databases.\n### 3. Build Agent-Ready Infrastructure: Perception and Action\nAI agents need infrastructure that supports two core capabilities: perceiving data and taking action on it.\n## Get Lak Lakshmanan’s stories in your inbox\nJoin Medium for free to get updates from this writer.\nSubscribe\nSubscribe\nRemember me for faster sign in\nNot all data formats are equally accessible to language model-based agents. Consider how easily an agent can parse, understand, and extract meaning from your data formats. Formats that preserve semantic meaning and require minimal preprocessing reduce friction.\nAI agents take action by calling tools — functions, APIs, and services that allow them to operate on data. Your infrastructure needs to make these tools discoverable and usable by agents. This means clear interfaces, good documentation, and reliable execution.\nPress enter or click to view image in full size\nSupporting the two core capabilities of an AI agent: perceiving data and taking action. Graphic generated by author using Gemini.\nAudit your data access patterns and tools from an AI agent’s perspective. What would an autonomous system need to know to use them effectively? Where is friction preventing smooth operation?\n### 4. Manage Agent Artifacts as First-Class Data\nAI agents don’t just consume data, they produce it. In fact, one of the things you will realize is that AI-generated content will start to greatly outweigh the “raw” data in your systems.\nWhen an agent generates output, makes a decision, writes code, or logs its reasoning process, that becomes data too.\nPress enter or click to view image in full size\nWhether content is created by humans, collected from software systems, or generated by AI agents, it will have to conform to the expectations and regulations that are common in your industry. Besides regulatory compliance, this agent-generated data has value for debugging, auditing, training future agents, and understanding system behavior.\nTreat agent artifacts with the same rigor as other data:\n * Store agent outputs systematically\n * Preserve decision logs and reasoning traces\n * Manage agent-generated code as versioned artifacts\n * Make this data accessible for analysis and future training\n\n\nThese artifacts become part of your data ecosystem. Design storage and access patterns accordingly.\n### 5. Connect Observation to Training\nThe fastest path to better agents is closing the loop between observability and training. AI agent infrastructure needs bidirectional pipelines that tie model performance and observability back into continuous training.\nFirst, you need an observability platform that tracks data quality metrics and, crucially, detects data drift (changes in input data characteristics) and concept drift (changes in the relationship between input and output). Simultaneously, it must monitor key model performance indicators such as accuracy, latency, and hallucination rates. Establish automated triggers tied to predefined thresholds.\nYour observability platform will also have to expand, to include human feedback. Every correction that your users make to generated content needs to be logged, and used to improve the AI models.\nPress enter or click to view image in full size\nTie observability and retraining. Graphic generated by author using Gemini\nSecond, you need a retraining pipeline that activates upon receiving the trigger event from the monitoring system. It must be fully automated, responsible for pulling the latest versioned and curated training data, initiating a model retraining or fine-tuning job, and running the newly trained model through a comprehensive suite of evaluation and regression tests. The construction of this closed-loop system, connecting performance monitoring directly to automated deployment, is a critical task for ML/data engineers (the boundaries will increasingly blur) in the agentic era.\n### Summary: How the role of a data engineer changes\nThese five shifts share a common theme: moving from rigid, predetermined workflows to flexible, context-aware infrastructure. Tool calling and reflection capabilities in modern agents make rigid ETL/ELT pipelines less relevant. In-context learning makes exemplar curation more valuable than exhaustive collection.\nPress enter or click to view image in full size\nHow the role of a data engineer changes in this new agentic era.\nData engineering isn’t becoming less important, it’s becoming different. The skills that built the data infrastructure of the last decade remain valuable, but they need to be applied toward different goals. Instead of designing every workflow in advance, you’re creating environments where agents can design workflows themselves.\n[Agentic Ai](https://medium.com/tag/agentic-ai?source=post_page---footer_tags--518861a8f025-----------------------------------------)\n[Data Engineering](https://medium.com/tag/data-engineering?source=post_page---footer_tags--518861a8f025-----------------------------------------)\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fvote%2Fai-advances%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&user=Lak+Lakshmanan&userId=247b0630b5d6&source=---footer_actions--518861a8f025---------------------clap_footer--------------------)\n--\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fvote%2Fai-advances%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&user=Lak+Lakshmanan&userId=247b0630b5d6&source=---footer_actions--518861a8f025---------------------clap_footer--------------------)\n--\n12\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Frepost%2Fp%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&user=Lak+Lakshmanan&userId=247b0630b5d6&source=---footer_actions--518861a8f025---------------------repost_footer--------------------)\n[](https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fbookmark%2Fp%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---footer_actions--518861a8f025---------------------bookmark_footer--------------------)\n[![Lak Lakshmanan](https://miro.medium.com/v2/resize:fill:96:96/2*TveVoapl-TEk-jBTrbis8w.jpeg)](https://lakshmanok.medium.com/?source=post_page---post_author_info--518861a8f025-----------------------------------------)\n[![Lak Lakshmanan](https://miro.medium.com/v2/resize:fill:128:128/2*TveVoapl-TEk-jBTrbis8w.jpeg)](https://lakshmanok.medium.com/?source=post_page---post_author_info--518861a8f025-----------------------------------------)\n## [Written by Lak Lakshmanan](https://lakshmanok.medium.com/?source=post_page---post_author_info--518861a8f025-----------------------------------------)\n[11.3K followers](https://lakshmanok.medium.com/followers?source=post_page---post_author_info--518861a8f025-----------------------------------------)\n·[89 following](https://medium.com/@lakshmanok/following?source=post_page---post_author_info--518861a8f025-----------------------------------------)\nCo-founder and CTO of [Obin.ai](http://Obin.ai) -- we build deep domain agents for finance. We are hiring. Please see open positions here: \n[![AI Advances](https://miro.medium.com/v2/resize:fill:96:96/1*R8zEd59FDf0l8Re94ImV0Q.png)](https://aiadvances.org/?source=post_page---post_publication_info--518861a8f025-----------------------------------------)\n[![AI Advances](https://miro.medium.com/v2/resize:fill:128:128/1*R8zEd59FDf0l8Re94ImV0Q.png)](https://aiadvances.org/?source=post_page---post_publication_info--518861a8f025-----------------------------------------)\n## [Published in AI Advances](https://aiadvances.org/?source=post_page---post_publication_info--518861a8f025-----------------------------------------)\n[111K followers](https://aiadvances.org/followers?source=post_page---post_publication_info--518861a8f025-----------------------------------------)\n·[Last published 1 hour ago](https://aiadvances.org/ai-agents-dont-need-bigger-context-windows-they-need-memory-db184f987306?source=post_page---post_publication_info--518861a8f025-----------------------------------------)\nIdeas in depth. Where exceptional thinking takes shape.\n[Help](https://help.medium.com/hc/en-us?source=post_page-----518861a8f025-----------------------------------------)\n[Status](https://status.medium.com/?source=post_page-----518861a8f025-----------------------------------------)\n[About](https://medium.com/about?autoplay=1&source=post_page-----518861a8f025-----------------------------------------)\n[Careers](https://medium.com/jobs-at-medium/work-at-medium-959d1a85284e?source=post_page-----518861a8f025-----------------------------------------)\nPress\n[Blog](https://blog.medium.com/?source=post_page-----518861a8f025-----------------------------------------)\n[Store](https://medium.com/store)\n[Privacy](https://policy.medium.com/medium-privacy-policy-f03bf92035c9?source=post_page-----518861a8f025-----------------------------------------)\n[Rules](https://policy.medium.com/medium-rules-30e5502c4eb4?source=post_page-----518861a8f025-----------------------------------------)\n[Terms](https://policy.medium.com/medium-terms-of-service-9db0094a1e0f?source=post_page-----518861a8f025-----------------------------------------)\n[Text to speech](https://speechify.com/medium?source=post_page-----518861a8f025-----------------------------------------)\n", "metadata": { "title": "Medium", "description": "What it means to get your data ready for AI (or) How AI Agents are changing the job of a Data Engineer I’m giving the keynote at the “Data Engineering in the Age of AI” conference on Nov. 13 …", "keywords": null, "author": "Lak Lakshmanan", "og:site_name": "Medium", "og:type": "article", "og:title": "What it means to get your data ready for AI", "og:description": "(or) How AI Agents are changing the job of a Data Engineer", "og:url": "https://aiadvances.org/what-it-means-to-get-your-data-ready-for-ai-518861a8f025", "og:image": "https://miro.medium.com/v2/resize:fit:877/1*1uXpjinYzN15cY56YbElHw.png", "twitter:app:name:iphone": "Medium", "twitter:app:id:iphone": "828256236", "twitter:site": "@AIAdvances", "twitter:app:url:iphone": "medium://p/518861a8f025", "twitter:image:src": "https://miro.medium.com/v2/resize:fit:877/1*1uXpjinYzN15cY56YbElHw.png", "twitter:card": "summary_large_image", "twitter:label1": "Reading time", "twitter:data1": "7 min read", "article:published_time": "2025-11-06T19:07:13.480Z", "article:author": "https://lakshmanok.medium.com" }, "paywall": { "listed_paid": true, "detected": false, "marker": "", "public_only": true, "policy": "付费/会员内容只保留页面公开可见部分，不绕过付费墙" }, "crawl": { "success": true, "status_code": 307, "error": "" }, "code_blocks": [], "images": [ { "src": "https://miro.medium.com/v2/resize:fill:64:64/1*dmbNkD5D-u45r44go_cf0g.png", "alt": "Unknown user", "local_path": "attachments/4036ef3c3f6d/images/000.png", "bytes": 1474, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fill:76:76/1*R8zEd59FDf0l8Re94ImV0Q.png", "alt": "AI Advances", "local_path": "attachments/4036ef3c3f6d/images/001.png", "bytes": 6641, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fill:64:64/2*TveVoapl-TEk-jBTrbis8w.jpeg", "alt": "Lak Lakshmanan", "local_path": "attachments/4036ef3c3f6d/images/002.jpeg", "bytes": 2153, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:640/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/003.png", "bytes": 21588, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:720/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/004.png", "bytes": 24946, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:750/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/005.png", "bytes": 26960, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:786/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/006.png", "bytes": 28554, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:828/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/007.png", "bytes": 30640, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1100/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/008.png", "bytes": 43194, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1400/format:webp/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/009.png", "bytes": 40610, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:640/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/010.png", "bytes": 147455, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:720/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/011.png", "bytes": 187562, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:750/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/012.png", "bytes": 212727, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:786/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/013.png", "bytes": 225685, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:828/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/014.png", "bytes": 235572, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1100/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/015.png", "bytes": 337004, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1400/1*9jZurC2oMPgobQBfPOSjgw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/016.png", "bytes": 184399, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:640/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/017.png", "bytes": 15078, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:720/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/018.png", "bytes": 16954, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:750/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/019.png", "bytes": 17276, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:786/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/020.png", "bytes": 18178, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:828/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/021.png", "bytes": 19194, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1100/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/022.png", "bytes": 25648, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1400/format:webp/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/023.png", "bytes": 33822, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:640/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/024.png", "bytes": 56026, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:720/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/025.png", "bytes": 64736, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:750/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/026.png", "bytes": 68006, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:786/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/027.png", "bytes": 72638, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:828/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/028.png", "bytes": 76916, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1100/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/029.png", "bytes": 106499, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1400/1*S95xyYvWcL1JNOly6hI7mg.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/030.png", "bytes": 143648, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:640/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/031.png", "bytes": 28136, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:720/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/032.png", "bytes": 30196, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:750/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/033.png", "bytes": 32018, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:786/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/034.png", "bytes": 34206, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:828/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/035.png", "bytes": 36562, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1100/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/036.png", "bytes": 36502, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:1400/format:webp/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/037.png", "bytes": 36502, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:640/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/038.png", "bytes": 105159, "downloaded": true }, { "src": "https://miro.medium.com/v2/resize:fit:720/1*1uXpjinYzN15cY56YbElHw.png", "alt": "", "local_path": "attachments/4036ef3c3f6d/images/039.png", "bytes": 122556, "downloaded": true } ], "audio": [], "videos": [], "links": [ { "type": "internal", "href": "https://aiadvances.org/sitemap/sitemap.xml" }, { "type": "internal", "href": "https://aiadvances.org/?source=post_page---publication_nav-3fe99b2acc4-518861a8f025-----------------------------------------" }, { "type": "internal", "href": "https://aiadvances.org/?source=post_page-----518861a8f025-----------------------------------------" }, { "type": "internal", "href": "https://aiadvances.org/?source=post_page---post_publication_sidebar-3fe99b2acc4-518861a8f025-----------------------------------------" }, { "type": "internal", "href": "https://aiadvances.org/?source=post_page---post_publication_info--518861a8f025-----------------------------------------" }, { "type": "internal", "href": "https://aiadvances.org/followers?source=post_page---post_publication_info--518861a8f025-----------------------------------------" }, { "type": "internal", "href": "https://aiadvances.org/ai-agents-dont-need-bigger-context-windows-they-need-memory-db184f987306?source=post_page---post_publication_info--518861a8f025-----------------------------------------" }, { "type": "external", "href": "https://play.google.com/store/apps/details?id=com.medium.reader&referrer=utm_source%3DmobileNavBar&source=---top_nav_layout_nav-------------------------------------------" }, { "type": "external", "href": "https://medium.com/m/signin?operation=login&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=post_page---top_nav_layout_nav-----------------------global_nav--------------------" }, { "type": "external", "href": "https://medium.com/?source=---top_nav_layout_nav-------------------------------------------" }, { "type": "external", "href": "https://medium.com/m/signin?operation=register&redirect=https%3A%2F%2Fmedium.com%2Fnew-story&source=---top_nav_layout_nav-----------------------new_post_topnav--------------------" }, { "type": "external", "href": "https://medium.com/search?source=---top_nav_layout_nav-------------------------------------------" }, { "type": "external", "href": "https://medium.com/tag/agentic-ai?source=post_page---header_tags--518861a8f025-----------------------------------------" }, { "type": "external", "href": "https://medium.com/tag/data-engineering?source=post_page---header_tags--518861a8f025-----------------------------------------" }, { "type": "external", "href": "https://lakshmanok.medium.com/?source=post_page---byline--518861a8f025-----------------------------------------" }, { "type": "external", "href": "https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fvote%2Fai-advances%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---header_actions--518861a8f025---------------------clap_footer--------------------&user=Lak+Lakshmanan&userId=247b0630b5d6" }, { "type": "external", "href": "https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Frepost%2Fp%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---header_actions--518861a8f025---------------------repost_header--------------------&user=Lak+Lakshmanan&userId=247b0630b5d6" }, { "type": "external", "href": "https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fbookmark%2Fp%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---header_actions--518861a8f025---------------------bookmark_footer--------------------" }, { "type": "external", "href": "https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2Fplans%3Fdimension%3Dpost_audio_button%26postId%3D518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---header_actions--518861a8f025---------------------post_audio_button--------------------" }, { "type": "external", "href": "https://www.oreilly.com/live-events/data-superstream-data-engineering-in-the-age-of-ai/0642572221676/0642572221669/" }, { "type": "external", "href": "https://github.com/CopilotKit/CopilotKit" }, { "type": "external", "href": "https://arxiv.org/abs/2201.11903" }, { "type": "external", "href": "https://www.promptingguide.ai/techniques/fewshot" }, { "type": "external", "href": "https://docs.cloud.google.com/agent-builder/agent-engine/example-store/overview" }, { "type": "external", "href": "https://medium.com/tag/agentic-ai?source=post_page---footer_tags--518861a8f025-----------------------------------------" }, { "type": "external", "href": "https://medium.com/tag/data-engineering?source=post_page---footer_tags--518861a8f025-----------------------------------------" }, { "type": "external", "href": "https://medium.com/m/signin?actionUrl=https%3A%2F%2Fmedium.com%2F_%2Fvote%2Fai-advances%2F518861a8f025&operation=register&redirect=https%3A%2F%2Faiadvances.org%2Fwhat-it-means-to-get-your-data-ready-for-ai-518861a8f025&source=---footer_actions--518861a8f025---------------------clap_footer--------------------&user=Lak+Lakshmanan&userId=247b0630b5d6" } ], "repair_version": 2, "repaired_at": "2026-10-14T11:11:49Z", "source_url": "https://aiadvances.org/what-it-means-to-get-your-data-ready-for-ai-518861a8f025", "delivery_status": "excluded_paid", "discovery_status": "excluded_paid", "missing_reason": "按原约定排除付费/会员全文" }