Stop Using LLMs Like Giant Problem Solvers

*   Cookie cookieyes-consent 
*   Duration 1 year 
*   Description CookieYes 设置此 Cookie 以记住用户的同意偏好，以便在后续访问此网站时尊重他们的偏好。它不会收集或存储有关网站访客的任何个人信息。 

Functional

- [x] 

功能性 Cookie 帮助执行某些功能，例如在社交媒体平台上分享网站内容、收集反馈和其他第三方功能。

*   Cookie lidc 
*   Duration 1 day 
*   Description LinkedIn 设置 lidc Cookie 以帮助选择数据中心。 

*   Cookie brw 
*   Duration 1 year 
*   Description 无描述可用。 

*   Cookie brwConsent 
*   Duration 5 minutes 
*   Description 描述目前不可用。 

*   Cookie WMF-Uniq 
*   Duration 1 year 
*   Description 描述目前不可用。 

*   Cookie loom_anon_comment 
*   Duration 1 year 
*   Description 无描述可用。 

*   Cookie loom_referral_video 
*   Duration session 
*   Description 描述目前不可用。 

*   Cookie VISITOR_INFO1_LIVE 
*   Duration 6 months 
*   Description 由 YouTube 设置的 Cookie，用于测量带宽，确定用户是否获得新旧播放器界面。 

*   Cookie yt-remote-connected-devices 
*   Duration Never Expires 
*   Description YouTube 设置此 Cookie 以使用嵌入的 YouTube 视频存储用户的视频偏好。 

*   Cookie ytidb::LAST_RESULT_ENTRY_KEY 
*   Duration Never Expires 
*   Description ytidb::LAST_RESULT_ENTRY_KEY Cookie 由 YouTube 使用，以存储用户点击的最后一项搜索结果。该信息用于通过提供更相关的搜索结果来改善用户体验。 

*   Cookie yt-remote-device-id 
*   Duration Never Expires 
*   Description YouTube 设置此 Cookie 以使用嵌入的 YouTube 视频存储用户的视频偏好。 

*   Cookie yt-remote-session-name 
*   Duration session 
*   Description yt-remote-session-name Cookie 由 YouTube 使用，以使用嵌入的 YouTube 视频存储用户的视频播放器偏好。 

*   Cookie yt-remote-fast-check-period 
*   Duration session 
*   Description yt-remote-fast-check-period Cookie 由 YouTube 使用，以存储嵌入的 YouTube 视频的用户视频播放器偏好。 

*   Cookie yt-remote-session-app 
*   Duration session 
*   Description yt-remote-session-app Cookie 由 YouTube 使用，以存储嵌入的 YouTube 视频播放器的用户偏好和界面信息。 

*   Cookie yt-remote-cast-available 
*   Duration session 
*   Description yt-remote-cast-available Cookie 用于存储用户关于其 YouTube 视频播放器上是否支持投射的偏好。 

*   Cookie yt-remote-cast-installed 
*   Duration session 
*   Description yt-remote-cast-installed Cookie 用于存储嵌入的 YouTube 视频的用户视频播放器偏好。 

*   Cookie cp_session 
*   Duration 3 months 
*   Description Codepen 为此网站上的帮助系统设置此 Cookie。 

*   Cookie loid 
*   Duration 1 year 1 month 4 days 
*   Description 此 Cookie 由 Reddit 设置。该 Cookie 启用将网站内容共享到社交媒体平台的功能。 

Analytics

- [x] 

分析性 Cookie 用于了解访客如何与网站互动。这些 Cookie 可以为访问量、跳出率、流量来源等指标提供信息。

*   Cookie __hstc 
*   Duration 6 months 
*   Description Hubspot 设置此主要 Cookie 以跟踪访客。它包含域名、初始时间戳（首次访问）、最后时间戳（最后一次访问）、当前时间戳（本次访问）以及会话编号（每次后续会话递增）。 

*   Cookie hubspotutk 
*   Duration 6 months 
*   Description HubSpot 设置此 Cookie 以跟踪网站的访客。此 Cookie 在表单提交时传递给 HubSpot，并在去重联系人时使用。 

*   Cookie _ga 
*   Duration 1 year 1 month 4 days 
*   Description Google Analytics 设置此 Cookie 以计算访客、会话和活动数据，并跟踪站点的使用情况以供站点的分析报告使用。该 Cookie 匿名存储信息并分配一个随机生成的数字以识别唯一访客。 

*   Cookie _ga_* 
*   Duration 1 year 1 month 4 days 
*   Description Google Analytics 设置此 Cookie 以存储和计数页面浏览次数。 

*   Cookie __Host-psifi.analyticsTrace 
*   Duration 6 hours 
*   Description 描述目前不可用。 

*   Cookie __Host-psifi.analyticsTraceV2 
*   Duration 6 hours 
*   Description 描述目前不可用。 

*   Cookie _gh_sess 
*   Duration session 
*   Description GitHub 设置此 Cookie 以在页面间临时存储应用程序和框架状态，例如用户在多步骤表单中的当前位置。 

*   Cookie YSC 
*   Duration session 
*   Description YSC Cookie 由 Youtube 设置，用于跟踪 Youtube 页面中嵌入视频的观看次数。 

*   Cookie ajs_anonymous_id 
*   Duration 1 year 
*   Description 此 Cookie 由 Segment 设置，以跟踪访问特定网站的人数，通过跟踪他们是否之前访问过。 

*   Cookie vuid 
*   Duration 1 year 1 month 4 days 
*   Description Vimeo 安装此 Cookie 通过设置唯一的 ID 来收集跟踪信息，以在网站上嵌入视频。  

Performance

- [x] 

性能 Cookie 用于理解并分析网站的关键性能指标，从而为访客提供更好的用户体验。

*   Cookie AWSALB 
*   Duration 7 days 
*   Description AWSALB 是 Amazon Web Services 设置的应用程序负载均衡器 Cookie，用于将会话映射到目标。 

*   Cookie acq 
*   Duration past 
*   Description 描述目前不可用。 

*   Cookie acq.sig 
*   Duration past 
*   Description Description is currently not available. 

*   Cookie ptc 
*   Duration 2 years 
*   Description No description available. 

Advertisement

- [x] 

Advertisement cookies are used to provide visitors with customised advertisements based on the pages you visited previously and to analyse the effectiveness of the ad campaigns.

*   Cookie muc_ads 
*   Duration 1 year 1 month 4 days 
*   Description Twitter sets this cookie to collect user behaviour and interaction data to optimize the website. 

*   Cookie guest_id_marketing 
*   Duration 1 year 1 month 4 days 
*   Description Twitter sets this cookie to identify and track the website visitor. 

*   Cookie guest_id_ads 
*   Duration 1 year 1 month 4 days 
*   Description Twitter sets this cookie to identify and track the website visitor. 

*   Cookie personalization_id 
*   Duration 1 year 1 month 4 days 
*   Description Twitter sets this cookie to integrate and share features for social media and also store information about how the user uses the website, for tracking and targeting. 

*   Cookie guest_id 
*   Duration 1 year 1 month 4 days 
*   Description Twitter sets this cookie to identify and track the website visitor. It registers if a user is signed in to the Twitter platform and collects information about ad preferences. 

*   Cookie bcookie 
*   Duration 1 year 
*   Description LinkedIn sets this cookie from LinkedIn share buttons and ad tags to recognize browser IDs. 

*   Cookie __Secure-ROLLOUT_TOKEN 
*   Duration 6 months 
*   Description YouTube sets this cookie to manage feature rollout and experimentation. It helps Google control which new features or interface changes are shown to users as part of testing and staged rollouts, ensuring consistent experience for a given user during an experiment. 

*   Cookie yt.innertube::nextId 
*   Duration Never Expires 
*   Description YouTube sets this cookie to register a unique ID to store data on what videos from YouTube the user has seen. 

*   Cookie yt.innertube::requests 
*   Duration Never Expires 
*   Description YouTube sets this cookie to register a unique ID to store data on what videos from YouTube the user has seen. 

*   Cookie session_tracker 
*   Duration session 
*   Description This cookie is set by the Reddit. This cookie is used to identify trusted web traffic. It also helps in adverstising on the website. 

*   Cookie edgebucket 
*   Duration session 
*   Description Reddit sets this cookie to save the information about a log-on Reddit user, for the purpose of advertisement recommendations and updating the content. 

*   Cookie did 
*   Duration 1 year 
*   Description Arbor sets this cookie to show targeted ads to site visitors.This cookie expires after 2 months or 1 year. 

Uncategorised

Other uncategorised cookies are those that are being analysed and have not been classified into a category as yet.

No cookies to display.

Reject All Save My Preferences Accept All

[Skip to content](https://towardsdatascience.com/stop-using-llms-like-giant-problem-solvers/#wp--skip-link--target)

[![Image 4: Towards Data Science](https://towardsdatascience.com/wp-content/uploads/2025/02/TDS-Vector-Logo.svg)](https://towardsdatascience.com/)

Publish AI, ML & data-science insights to a global community of data professionals.

[Sign in](https://towardsdatascience.com/stop-using-llms-like-giant-problem-solvers/)

[Submit an Article](https://contributor.insightmediagroup.io/)

*   [Latest](https://towardsdatascience.com/latest/)
*   [Editor’s Picks](https://towardsdatascience.com/tag/editors-pick/)
*   [Deep Dives](https://towardsdatascience.com/tag/deep-dives/)
*   [Newsletter](https://towardsdatascience.com/tag/the-variable/)

* * *

*   [Write For TDS](https://towardsdatascience.com/submissions/)

[![Image 5: Towards Data Science](https://towardsdatascience.com/wp-content/uploads/2025/02/TDS-Vector-Logo.svg)](https://towardsdatascience.com/)

Toggle Mobile Navigation

*   [LinkedIn](https://www.linkedin.com/company/towards-data-science/?originalSubdomain=ca)
*   [X](https://x.com/TDataScience)

Toggle Search

Search 

[LLM Applications](https://towardsdatascience.com/category/artificial-intelligence/llm-applications/)

# Stop Using LLMs Like Giant Problem Solvers

How I turned 100 messy pdfs into structured insights by building a deterministic loop around agents

[Clara Chong](https://towardsdatascience.com/author/clarachong13/)

May 26, 2026

 6 min read

 Share 

![Image 6](https://towardsdatascience.com/wp-content/uploads/2026/05/documents-medium-scaled-1.jpg)

[Image](https://unsplash.com/photos/stacks-of-paper-documents-and-file-folders-snNHKZ-mGfE) by Wesley Tingey from Unsplash

I recently worked on a feature where I had to transform 100 messy compliance pdfs into structured JSON rules.

The brute force approach was obvious: give the agent the source text, explain the task, provide examples, and ask it to generate the rules. Since it was the lowest-hanging fruit, I tried it first.

At a glance, the output looked fine. The output JSON was valid and matched what I expected.

But as I was manually sampling the results to check for accuracy, the cracks appeared. Some rules were too broad, others were missed. Some rules failed to preserve the nuances of the original text. I tried using another agent to catch and fix the errors but with such a huge corpus, it was impossible to confidently verify the output.

That was the frustrating part. The errors were not obvious. This was way too fragile of an implementation to scale.

Though I cannot share the exact implementation details, what I can share are the architectural lessons I learnt and how I eventually implemented it. Hopefully, these insights will be useful if you’re building AI systems that need to scale, stay reliable, and deal with messy data. And if you have better ways of doing things, [do reach out to chat](https://www.linkedin.com/in/clarachongjy/)!

## The problem

The 100 PDFs I worked with had already been parsed and chunked before they reached me. But the raw content was still messy. There were bullet points, tables, OCR artifacts, translated sections, semi-structured headings, footers, headers, inconsistent formatting, and document-specific quirks.

I chose to use an agent because deciding what mattered required semantic judgment. The documents did not follow one consistent pattern, so relevance could not be determined through simple rules alone.

You had to understand the surrounding context. None of this was difficult when done on a small chunk of data. **The challenge was performing this reliably at scale.**

These rules were then processed by another downstream system to be evaluated deterministically.

## What eventually worked

After a few experiments, I realized the biggest improvement did not come from a better prompt, a new tool, an MCP server, or a more sophisticated agent harness.

It came from changing the shape of the problem.

> **Instead of trying to make the agent smarter, I made the agent's job smaller.**

The first change was to prepare the source data upfront. Instead of asking the agent to query a database, retrieve records, decide whether it had the right inputs, and then perform the extraction, I gave it a more controlled starting point.

In my case, that meant temporarily storing the relevant raw data locally.

This may not always be practical. But the underlying principle is to reduce the amount of retrieval uncertainty the agent has to handle. If the agent's job is to reason over content, don't also make it responsible for figuring out whether it has found the right content.

Another option would be to prepare the query upfront.

I also used a script to strip away unnecessary metadata and fields before passing the raw content to the agent. Less irrelevant context meant fewer distractions, fewer chances for the agent to latch onto the wrong details, and a cleaner reasoning task overall.

> **But the most important change was the unit of work.**
> 
> Instead of processing everything at once, I did things iteratively and processed one document at a time.

That made each job smaller, easier to inspect, easier to retry, and easier to audit. I spun up five sub-agents to process documents in parallel, with each agent logging its progress to a file.

If one document failed, I could retry only that document. If one output had formatting issues, I could fix that specific case without rerunning the whole batch. If the pipeline stopped halfway, the cached progress meant it could resume from the last successful checkpoint.

This was also where the separation of responsibilities became clearer.

The agent handled the semantic work: understanding the content, identifying the relevant parts, and writing the JSON output.

The surrounding code handled the mechanical parts: parallelizing jobs, enforcing the schema, generating IDs, writing files, caching progress, validating references, and checking whether the output could be traced back to the original source.

I also had an orchestrator watch over the progress of the script.

## Making the output auditable

A useful design decision was adding reference IDs to every generated rule. This meant that each output item pointed back to a specific source.

This made the output easier to audit. Instead of asking, "Does this generated rule look right?" I could ask more precise questions such as: does the referenced source chunk exist? Is the quoted source text actually present in that chunk?

I could also get another agent to selectively run audits on larger and more complex documents to ensure that important nuances were preserved.

On top of that, I did a lightweight version of evaluations. I ran a small batch of raw documents through the workflow and manually reviewed the results for coverage and accuracy. A full gold dataset was not practical for the scope of this task, but I still needed a way to prove to myself that the workflow was working.

My goal was not to build a perfect benchmark but to make the system auditable enough that I could inspect the outputs, catch failures, and iterate toward a higher accuracy bar.

If you've got ideas on how I could have done this better, let me know!

## My biggest takeaway

The pattern that worked was to stop treating the LLM as the whole system.

> The system became more reliable not because the agent became perfect, but because the workflow made its outputs easier to trace, validate, and recover from.

Coincidentally, I was building this shortly before attending the inaugural AI Engineer Singapore conference, held from 15–17 May 2026.

On the last day, JJ Geewax, Director of Applied AI at Google DeepMind, shared a framing that captured what I had been learning the hard way: **we need to stop using LLMs like giant problem solvers.**

That resonated with me because it is such an easy trap to fall into. It is easy to just give the model the data, schema, business rules, edge cases, and the responsibility to verify itself. Then get frustrated when the result is inconsistent.

But for reliable production systems, the better pattern is usually a hybrid. Let the agent handle the parts that require semantic judgment, and let code handle the parts that require structure, validation, and control.

I'll be sharing more reflections from AI Engineer Singapore and the workshops I attended. [The YouTube snippet of JJ's speech here](https://www.youtube.com/live/m12vGjfbNlo?t=5563s).

That's all from me. I hope this helped, and see you in the next article! [Image 7: 🙂](https://s.w.org/images/core/emoji/17.0.2/svg/1f642.svg)

* * *

Written By

Clara Chong

[See all from Clara Chong](https://towardsdatascience.com/author/clarachong13/)

[Agentic Ai](https://towardsdatascience.com/tag/agentic-ai/), [Artificial Intelligence](https://towardsdatascience.com/tag/artificial-intelligence/), [Editors Pick](https://towardsdatascience.com/tag/editors-pick/), [Llm](https://towardsdatascience.com/tag/llm/), [Programming](https://towardsdatascience.com/tag/programming/)

Share This Article

*   [Share on Facebook](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Ftowardsdatascience.com%2Fstop-using-llms-like-giant-problem-solvers%2F&title=Stop%20Using%20LLMs%20Like%20Giant%20Problem%C2%A0Solvers)
*   [Share on LinkedIn](https://www.linkedin.com/shareArticle?mini=true&url=https%3A%2F%2Ftowardsdatascience.com%2Fstop-using-llms-like-giant-problem-solvers%2F&title=Stop%20Using%20LLMs%20Like%20Giant%20Problem%C2%A0Solvers)
*   [Share on X](https://x.com/share?url=https%3A%2F%2Ftowardsdatascience.com%2Fstop-using-llms-like-giant-problem-solvers%2F&text=Stop%20Using%20LLMs%20Like%20Giant%20Problem%C2%A0Solvers)

Towards Data Science is a community publication. Submit your insights to reach our global audience and earn through the TDS Author Payment Program.

[Write for TDS](https://towardsdatascience.com/questions-96667b06af5/)

## Related Articles

*   ![Image 8](https://towardsdatascience.com/wp-content/uploads/2025/06/Grammar-as-an-Injectable-A-Trojan-Horse-to-NLP-How-machines-make-sense-of-sentence-structure-Combinatory-Categorial-Grammar-by-marco-hening-tallarico.webp)## [Grammar as an Injectable: A Trojan Horse to NLP](https://towardsdatascience.com/grammar-as-a-trojan-horse-to-nlp-and-computer-science/)

[Natural Language Processing](https://towardsdatascience.com/category/artificial-intelligence/nlp/) How machines make sense of sentence structure: Combinatory Categorial Grammar [Marco Hening Tallarico](https://towardsdatascience.com/author/marco-heningtallarico/) June 2, 2025  12 min read      
*   ![Image 9](https://towardsdatascience.com/wp-content/uploads/2025/07/doppleware_two_modern_robots_using_a_tin-can_phone_and_wire_t_bff835d3-b41a-4957-aeb4-13226f1760b2_0.png)## [Talk to my Agent](https://towardsdatascience.com/talk-to-my-agent/)

[Artificial Intelligence](https://towardsdatascience.com/category/artificial-intelligence/) The exciting new world of designing conversation-driven APIs for LLMs. [Roni Dover](https://towardsdatascience.com/author/roni-dover/) July 28, 2025  9 min read      
*   ![Image 10: Illustration of robots evaluating a model](https://towardsdatascience.com/wp-content/uploads/2025/07/image-345.png)## [How to Evaluate Graph Retrieval in MCP Agentic Systems](https://towardsdatascience.com/evaluating-graph-retrieval-in-mcp-agentic-systems/)

[Artificial Intelligence](https://towardsdatascience.com/category/artificial-intelligence/) A framework for measuring retrieval quality in Model Context Protocol agents. [Tomaz Bratanic](https://towardsdatascience.com/author/bratanic-tomaz/) July 29, 2025  9 min read      
*   ![Image 11](https://towardsdatascience.com/wp-content/uploads/2025/08/ChatGPT-Image-Aug-3-2025-11_57_46-AM.png)## [Finding Golden Examples: A Smarter Approach to In-Context Learning](https://towardsdatascience.com/finding-golden-examples-a-smarter-approach-to-in-context-learning/)

[LLM Applications](https://towardsdatascience.com/category/artificial-intelligence/llm-applications/) From random example selection to systematic AuPair generation—how to make your LLM prompts actually… [Sudheer Singh](https://towardsdatascience.com/author/sudheer-singh/) August 7, 2025  7 min read      
*   ![Image 12](https://towardsdatascience.com/wp-content/uploads/2025/08/ChatGPT-Image-20-ago-2025-09_29_42.jpg)## [“Where’s Marta?”: How We Removed Uncertainty From AI Reasoning](https://towardsdatascience.com/interactive-proofs-with-claude/)

[LLM Applications](https://towardsdatascience.com/category/artificial-intelligence/llm-applications/) A primer on overcoming LLM limitations with formal verification. [Jacopo Tagliabue](https://towardsdatascience.com/author/jacopotagliabue/) August 20, 2025  12 min read      
*   ![Image 13](https://towardsdatascience.com/wp-content/uploads/2025/08/unlocking-multimodal-video-transcription.gif)## [Unlocking Multimodal Video Transcription with Gemini](https://towardsdatascience.com/unlocking-multimodal-video-transcription-with-gemini/)

[LLM Applications](https://towardsdatascience.com/category/artificial-intelligence/llm-applications/) Explore how to transcribe videos with speaker identification in a single prompt [Laurent Picard](https://towardsdatascience.com/author/laurent-picard/) August 29, 2025  66 min read      
*   ![Image 14](https://towardsdatascience.com/wp-content/uploads/2025/08/0Ul9PApxHSz02D3_D.webp)## [How to Develop a Bilingual Voice Assistant](https://towardsdatascience.com/developing-a-bilingual-voice-assistant/)

[LLM Applications](https://towardsdatascience.com/category/artificial-intelligence/llm-applications/) Exploring ways to make voice assistants more personal [Deepak Krishnamurthy](https://towardsdatascience.com/author/deepak-krish/) August 31, 2025  8 min read      

*   [YouTube](https://www.youtube.com/c/TowardsDataScience)
*   [X](https://x.com/TDataScience)
*   [LinkedIn](https://www.linkedin.com/company/towards-data-science/?originalSubdomain=ca)
*   [Threads](https://www.threads.net/@towardsdatascience)
*   [Bluesky](https://bsky.app/profile/towardsdatascience.com)

[![Image 15: Towards Data Science](https://towardsdatascience.com/wp-content/uploads/2025/02/TDS-Vector-Logo.svg)](https://towardsdatascience.com/)

Your home for data science and AI. The world’s leading publication for data science, data analytics, data engineering, machine learning, and artificial intelligence professionals.

© Insight Media Group, LLC 2026 

Subscribe to Our Newsletter