# Large Language Model Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/large-language-model
> Markdown URL: https://aitinkerers.org/technologies/large-language-model.md
> Technology record last updated: 2026-02-23T14:19:32Z
> Generated: 2026-09-22T12:38:10Z

Large Language Models (LLMs) are transformer-based neural networks, trained on massive text data, that predict and generate human-quality language and code.

LLMs are deep learning models (e.g., GPT-4, Llama 3) built on the transformer architecture, which uses a self-attention mechanism to process massive, diverse datasets . These models, containing billions to trillions of parameters, function as general-purpose sequence predictors: they calculate the most statistically probable next token in a sequence . This core capability enables diverse applications, including summarization, code generation, translation, and complex reasoning (chain-of-thought) . While powerful, LLMs require significant compute resources and carry risks like generating false information (hallucinations) .

- Official technology site: https://en.wikipedia.org/wiki/Large_language_model
- Public AI Tinkerers demos and talks: 9
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Automated context compaction strategies](https://san-diego.aitinkerers.org/talks/rsvp_7lJPz-RERho)

This talk demonstrates how leveraging concepts from database design to manage context and semantically indexed time-weighted memories can improve the quality of interactions with LLMs while reducing token usage. It introduces a new concept called Context Structured Merge that only retrieves the most relevant historical context on demand, reducing token usage over the lifetime of a model interaction.

- Event context: Self-hosting Models and Managing Token Spend — 2026-08-21 — San Diego
- Public talk page: https://san-diego.aitinkerers.org/talks/rsvp_7lJPz-RERho

### [Turning real business SOPs into rules an agent can actually follow](https://da-nang.aitinkerers.org/talks/rsvp__ey6nXp67PY)

We built a harness for AI agents that execute real-world business operations from human-written SOPs. The system separates execution into three planes: Policy turns natural-language operating procedures into goals, guidance, authority boundaries and guardrails; Execution plans and coordinates long-running work across AI, humans and tools while adapting to changing real-world state; and Quality evaluates execution against the governing standard and operational evidence. Live, I’ll show the working system from SOP interpretation through adaptive execution and run evaluation.

- Event context: AI Tinkerers Da Nang: Inaugural Meetup — 2026-08-15 — Da Nang
- Public talk page: https://da-nang.aitinkerers.org/talks/rsvp__ey6nXp67PY

### [AI Boost - a context sharing system](https://valencia.aitinkerers.org/talks/rsvp_Zd-yr9DNsxM)

AI Boost is a context sharing MCP that allows repeatedly used context to be easily searched for and added to the LLM

- Event context: AI Tinkerers Valencia July Demo Night — 2026-07-28 — Valencia
- Public talk page: https://valencia.aitinkerers.org/talks/rsvp_Zd-yr9DNsxM

### [Data Entry automation using Browser Agents](https://dubai.aitinkerers.org/talks/rsvp_Capdf8Q-Kyc)

An AI browser agent that reads a weekly timesheet from Excel and automatically logs every hour into our system no clicks, no copy-paste, no manual entry. In the demo, you'll watch it open a live Salesforce environment, navigate to the correct date, intelligently match projects and tasks, and enter the hours in real time all driven by Claude reading and interacting with the screen just like a human would so that you can focus ur time on more important stuff.

- Event context: AI Tinkerers Dubai - June Demo Day — 2026-06-13 — Dubai
- Public talk page: https://dubai.aitinkerers.org/talks/rsvp_Capdf8Q-Kyc

### [Marketing Campaign Automation with Claude Code](https://seattle.aitinkerers.org/talks/rsvp_o38nCKHKmwI)

There are many steps to creating an effective marketing campaign, from strategic goals to asset/copy generation, execution, and tracking. For the conference I'm organizing, my goal was to create about 40 campaigns in 80 days. A daunting task. Unless you automate most of the research, planning, copy generation, project management to Claude Code.

- Event context: AI Tinkerers Seattle: GTM Track — March — 2026-03-26 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_o38nCKHKmwI

### [Reliable ML Predictions with Conformal Prediction](https://edmonton.aitinkerers.org/talks/rsvp_pjcUgYgII2c)

ML models make prediction but these predictions can be wrong. In high-stakes environments like healthcare, this has consequences. In normal everyday tasks like search or text generation, these wrong predictions are known as hallucinations. To integrate and augment ML models in society, we need them to be reliable! In this demo, I will show how to implement Conformal Prediction, a model-free framework that reasons how confident our model is. Ideally, if our model is not confident then, they should say "I don't know". However, current model deployments don't do this. Conformal Prediction wraps around any model to output a set instead of a single point. The sets are guaranteed to contain the true label with statistical validity. With this, we can statistically guarantee how reliable our predictions are instead of using adhoc unreliable methods that have no guarantees. I will be demo'ing how to convert your model to be more reliable with Conformal Prediction. This framework is general and useful in many applications. This will be a live demo that goes through the coding process and technical implementation of Conformal Prediction.

- Event context: AI Tinkerers Edmonton: March 2026 Meetup — 2026-03-25 — Edmonton
- Public talk page: https://edmonton.aitinkerers.org/talks/rsvp_pjcUgYgII2c

### [From Image to Structured Data: Building a Local AI Document OCR Platform for Administrative Workflows](https://tokyo.aitinkerers.org/talks/rsvp_Vs-o12h3f_U)

Administrative and compliance-heavy professions still rely heavily on paper and scanned documents. However, sending sensitive documents to cloud OCR or AI services is often not acceptable due to privacy, regulatory, or client confidentiality requirements. In this talk, I will present a professional web-based OCR processing platform designed for secure, local-first document handling — with a focus on real-world administrative document workflows such as those handled by 行政書士 professionals.

- Event context: AI Tinkerers Tokyo - Toranomon Meetup - February 19, 2026 — 2026-02-19 — Tokyo
- Public talk page: https://tokyo.aitinkerers.org/talks/rsvp_Vs-o12h3f_U

### [Self-reflection in agentic workflows](https://zurich.aitinkerers.org/talks/rsvp_4jGx0MOTPXE)

My demo will explore self-reflection in agentic workflows. You will learn how structured self-reflection enhances agentic process quality and prevents system deviations. I will act as an agentic system that executes actions and an AI verifier will monitor my actions and assess my execution. Perhaps I will even be able to trick the verifier ;) Some more details about the demo: I will be impersonating a browser interaction agent (similar to Google's Project Mariner or OpenAI's Operator) and the verifier (based on an LLM with vision) will use natural language description of my actions and screenshots of the webpage as input for verification.

- Event context: AI Tinkerers Zurich - February 6 — 2025-02-06 — Zürich
- Public talk page: https://zurich.aitinkerers.org/talks/rsvp_4jGx0MOTPXE

### [Tackling ARC via LLM Program Search](https://singapore.aitinkerers.org/talks/rsvp_AQBmuOeOAIc)

This talk details the work-in-progress techniques to tackle the ARC challenge. ARC is a set of "IQ challenges" made for AI that's easy for humans to solve, but can't be solved by any LLM - even OpenAI o1!

- Event context: AI Tinkerers Singapore: Meetup - September 16th, 2024 — 2024-09-16 — Singapore
- Public talk page: https://singapore.aitinkerers.org/talks/rsvp_AQBmuOeOAIc

## Related Technologies

- [ABBYY FineReader](https://aitinkerers.org/technologies/abbyy-finereader) ([Markdown](https://aitinkerers.org/technologies/abbyy-finereader.md)) — 3 public demos
- [AI](https://aitinkerers.org/technologies/ai) ([Markdown](https://aitinkerers.org/technologies/ai.md)) — 55 public demos
- [AI models](https://aitinkerers.org/technologies/ai-models) ([Markdown](https://aitinkerers.org/technologies/ai-models.md)) — 6 public demos
- [Amazon Textract](https://aitinkerers.org/technologies/amazon-textract) ([Markdown](https://aitinkerers.org/technologies/amazon-textract.md)) — 5 public demos
- [ARC](https://aitinkerers.org/technologies/arc) ([Markdown](https://aitinkerers.org/technologies/arc.md)) — 1 public demo
- [Artificial Intelligence](https://aitinkerers.org/technologies/artificial-intelligence) ([Markdown](https://aitinkerers.org/technologies/artificial-intelligence.md)) — 5 public demos
- [BERT](https://aitinkerers.org/technologies/bert) ([Markdown](https://aitinkerers.org/technologies/bert.md)) — 179 public demos
- [Browser Use](https://aitinkerers.org/technologies/browser-use) ([Markdown](https://aitinkerers.org/technologies/browser-use.md)) — 6 public demos
- [Browser Use framework](https://aitinkerers.org/technologies/browser-use-framework) ([Markdown](https://aitinkerers.org/technologies/browser-use-framework.md)) — 1 public demo
- [Claude](https://aitinkerers.org/technologies/claude) ([Markdown](https://aitinkerers.org/technologies/claude.md)) — 174 public demos
- [claude-4](https://aitinkerers.org/technologies/claude-4) ([Markdown](https://aitinkerers.org/technologies/claude-4.md)) — 8 public demos
- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 216 public demos
- [Cloud Vision API](https://aitinkerers.org/technologies/cloud-vision-api) ([Markdown](https://aitinkerers.org/technologies/cloud-vision-api.md)) — 3 public demos
- [Computer Vision](https://aitinkerers.org/technologies/computer-vision) ([Markdown](https://aitinkerers.org/technologies/computer-vision.md)) — 22 public demos
- [Conformal Prediction](https://aitinkerers.org/technologies/conformal-prediction) ([Markdown](https://aitinkerers.org/technologies/conformal-prediction.md)) — 1 public demo
- [Context Structured Merge](https://aitinkerers.org/technologies/context-structured-merge) ([Markdown](https://aitinkerers.org/technologies/context-structured-merge.md)) — 1 public demo
- [CRM](https://aitinkerers.org/technologies/crm) ([Markdown](https://aitinkerers.org/technologies/crm.md)) — 4 public demos
- [Data](https://aitinkerers.org/technologies/data) ([Markdown](https://aitinkerers.org/technologies/data.md)) — 8 public demos