# Qwen Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/qwen
> Markdown URL: https://aitinkerers.org/technologies/qwen.md
> Technology record last updated: 2026-02-22T18:20:26Z
> Generated: 2026-09-21T16:46:17Z

Alibaba Cloud's Qwen is a family of advanced, multilingual large language and multimodal models (LLM/LMM) with both proprietary and open-weight versions.

Qwen is the large language model (LLM) and large multimodal model (LMM) family engineered by Alibaba Cloud, designed for state-of-the-art text, vision, and audio processing. The model series, including the Qwen3 generation, features a comprehensive range of dense and Mixture-of-Experts (MoE) models, scaling from the efficient Qwen3-0.6B to the powerful Qwen3-235B-A22B. Qwen excels in complex tasks: it supports reasoning, agent capabilities (tool use), and instruction-following across 119 languages, with open-weight variants available under the Apache 2.0 license for broad deployment.

- Official technology site: https://chat.qwen.ai
- Public AI Tinkerers demos and talks: 18
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Dig.rest](https://orange-county.aitinkerers.org/talks/rsvp_a0OCfbGWoFA)

Digging into events of the world, building cause and effect maps and getting to know the layers behind the news.

- Event context: AI Tinkerers Orange County: Tuesday, July 21, 2026 at Centercode — 2026-07-22 — Orange County
- Public talk page: https://orange-county.aitinkerers.org/talks/rsvp_a0OCfbGWoFA

### [Emotional Poker](https://prague.aitinkerers.org/talks/rsvp_q05KA6GeJo4)

What if we have an environment where agents play poker against one another. This is a fun showcase by itself, but I am playing around with emotional vectors now. Basically you can find a "dishonest" "risk-taking" "risk-awerse" internal vector and use it to change the behavior of the agents. Then you can experiment how these "emotions" affect the agents in how they play.

- Event context: AI Tinkerers Prague June Meetup — 2026-06-12 — Prague
- Public talk page: https://prague.aitinkerers.org/talks/rsvp_q05KA6GeJo4

### [OpenClaw on the Edge - Running on 8GB Nvidia Jetson](https://london.aitinkerers.org/talks/rsvp_OcgzwBCExTI)

Connecting OpenClaw to terminal agent feeding the output summary to telegram channel with Qwen3-4B. Discussing the technical limitations of the system and the edge factor as well as the custom routing scripts.

- Event context: OpenClaw Demo Night // 5th March — 2026-03-05 — London
- Public talk page: https://london.aitinkerers.org/talks/rsvp_OcgzwBCExTI

### [Viendo como pasas tu tiempo con IA](https://bogota.aitinkerers.org/talks/rsvp_VDCkqnqOLKQ)

Todos queremos dedicar nuestro tiempo a cosas que valen la pena, pero la pantalla rara vez refleja esa intención. Durante el día producimos un rastro de acciones (apps, ventanas, documentos, cambios de contexto) que cuenta una historia distinta a la que creemos estar viviendo. En esta charla presento una herramienta que estoy construyendo para cerrar esa brecha: transforma mi actividad diaria en un resumen interpretable de “en qué se fue el tiempo”, lo contrasta con lo que yo considero importante, y me ayuda a ajustar el día siguiente. La implementé con modelos locales para mantener los datos en mi máquina, reducir fricción y poder operar offline. También la construí apoyándome en varias herramientas de IA para acelerar diseño, prototipado y evaluación. Compartiré el enfoque técnico y aprendizajes prácticos sobre cómo instrumentar tu propio sistema de reflexión diaria sin sacrificar privacidad.

- Event context: AI Tinkerers Bogotá: El Primer Meetup de 2026 — 2026-02-26 — Bogotá
- Public talk page: https://bogota.aitinkerers.org/talks/rsvp_VDCkqnqOLKQ

### [Extracting RFC 5545 RRULE Compliant Schedule Data in valid JSON with only 0.6B Parameters](https://seattle.aitinkerers.org/talks/rsvp_DaYlLeV25ZU)

We have a vast amount of unstructured schedule data on community services available to communities across 20 states. It is easier for data managers and community members to write brief notes, though service delivery schedules can be quite complex when translated to data that is interoperable with the iCal standard. This makes it a great task for LLMs... but VRAM/RAM and compute is expensive, you know? I plan to probably show what a row of our data looks like, a community service offered for those in need and its unstructured schedule data. Then, I will take that unstructured data, prompt the model with it, and we can all see if it generated something useful. It will probably all be CLI but I will zoom in my screen ʕ•ᴥ•ʔ

- Event context: AI Tinkerers Seattle: January Meetup — 2026-01-31 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_DaYlLeV25ZU

### [LLMs for retrieval and recommendation](https://toronto.aitinkerers.org/talks/rsvp_fnoZbMMW_ao)

I will show how to use vLLM and an openweight model to make a simple recommendation engine and use guided decoding to limit the output of the llm to the allowed items only. No finetuning needed and it will work on google colab so basically no hardware needed either. the code i shared will be a bit more, that is just a draft.

- Event context: AI Tinkerers Toronto - October 2025 Meetup at CIBC | Simplii — 2025-10-30 — Toronto
- Public talk page: https://toronto.aitinkerers.org/talks/rsvp_fnoZbMMW_ao

### [From Slop to Storytelling - Creating Anime with AI](https://la.aitinkerers.org/talks/rsvp_oNNyZn72V0w)

Once you start combining AI videos into a longer sequence, many challenges arise: - consistent characters - style transfer - pacing - lighting - emotion - voice and SFX - implied frame rate ...and more! We will discuss common workflows, techniques, and trade-offs when creating longer-form AI video.

- Event context: AI Tinkerers LA – October 2025: Ghosts in the Machine w/ Oxen.ai — 2025-10-21 — Los Angeles
- Public talk page: https://la.aitinkerers.org/talks/rsvp_oNNyZn72V0w

### [Edge AI: exploring the capabilities of Apple’s VLM](https://raleigh.aitinkerers.org/talks/rsvp_55jgZ4RgKLY)

Apple has released a VLM that’s a quantized fine tuned version of Qwen they’ve optimized for iOS and macOS Apple Silicon devices. I want to show some experiments on when it works and when it fails. For example how good is it at Q&A? How responsive to prompting is it? What languages can it work with both visually and textually? Are resource usages different on different hardware? What tunability does Apple offer by default?

- Event context: AI Tinkerers - Raleigh Inaugural Meetup (September 2025) — 2025-09-30 — Raleigh
- Public talk page: https://raleigh.aitinkerers.org/talks/rsvp_55jgZ4RgKLY

### [News x Dilbert: Using AI to Curate Comics for Today's News](https://hong-kong.aitinkerers.org/talks/rsvp_uVr4OwbRNbI)

Matching today’s news headlines with the most relevant Dilbert comic strip from over 35 years of archives. This is a hobbyist project to showcase how AI technology can be applied.

- Event context: AI Tinkerers Hong Kong with AWS Meetup — September 29, 2025 — 2025-09-29 — Hong Kong
- Public talk page: https://hong-kong.aitinkerers.org/talks/rsvp_uVr4OwbRNbI

### [Parlina: plataforma de educación con IA agencial](https://manizales.aitinkerers.org/talks/rsvp_FT3SMEKIqUo)

Presento Parlina, una plataforma de aprendizaje con IA que convierte información (interna o vía Deep Research) en competencias verificables. Demo del flujo pedagógico: Taxonomía → Perfil del estudiante → Objetivos → generación multimodal (texto/imagen/video). En curso: generación e interacción por audio, control de calidad automático + grounding, y curaduría humana, métricas en el LMS y certificación basada en objetivos.

- Event context: 🚀 ¡Octavo Encuentro de AI Tinkerers Manizales! — 2025-09-24 — Manizales
- Public talk page: https://manizales.aitinkerers.org/talks/rsvp_FT3SMEKIqUo

### [SLM for MacOs](https://poland.aitinkerers.org/talks/rsvp_kDFlECrDcVM)

Challenges and opportunities of using SLM on MacOs platform

- Event context: AI Tinkerers Poland #5 - Meetup in Warsaw (September) — 2025-09-18 — Poland
- Public talk page: https://poland.aitinkerers.org/talks/rsvp_kDFlECrDcVM

### [Parsing complex pdfs using vision models](https://dubai.aitinkerers.org/talks/rsvp_DTChla5ieP8)

My experience converting complex PDFs (containing tables and math formulas) using lama4 maverick and Qwen 2.5 VI. I want to show the importance of image resolution in having good text/formulas/table extraction. Hopefully this will save some time for those trying to do the same thing

- Event context: AI Tinkerers Dubai Meetup – June 2025 Demo Day — 2025-06-28 — Dubai
- Public talk page: https://dubai.aitinkerers.org/talks/rsvp_DTChla5ieP8

### [Language Driven Organ-Lesion Predictions in a Large Scale Toxicological Experiment](https://milan.aitinkerers.org/talks/rsvp_SO2HacjeF2k)

Predicting compound toxicity relies predominantly on representing molecules at various levels of detail, with evaluation based on extensive experimental studies. These studies test a vast array of factors to statistically estimate the potentially harmful outcomes of treatment conditions. This research analyzes toxicology experiments from a new perspective. Rather than studying experimental factors separately or developing specialized deep learning models, this approach interprets experimental factors as strings and uses general-purpose pretrained language models for surprisingly accurate predictions. Textual descriptors are either projected into a dense vector space with embedding models and then integrated into a nested cross-validation pipeline, or given as input to state-of-the-art Large Language Models (LLMs) that directly attempt zero-shot classification using schema-constrained generation. The experimental validation uses data from the repeated-dose Open TG-GATEs dataset, which exposes the same rat clone to 142 compounds over four different time periods at three levels of dosage and contains histopathology annotations for both kidney and liver lesions.

- Event context: AI Tinkerers Milan - June 10, 2025 - Community Demos & Networking — 2025-06-10 — Milan
- Public talk page: https://milan.aitinkerers.org/talks/rsvp_SO2HacjeF2k

### [Fine-Tuning Models for Content Moderation with Apple’s MLX Framework](https://orange-county.aitinkerers.org/talks/rsvp_S0EeHEPZUv4)

This talk is a hands-on walkthrough of how to fine-tune large language models—and multi-modal models—locally on Apple-Silicon Macs using Apple’s open-source MLX framework. We’ll: 1. Give a quick MLX overview and why it’s optimized for the M-series GPU/ANE. 2. Explain LoRA / QLoRA and why adapter-based fine-tuning is memory-efficient. 3. Dissect the vision-encoder → adapter → LLM pipeline (SigLIP + Phi-1.5). 4. Show dataset prep in JSONL, CLI commands (mlx_lm.lora, mlx_lm.fuse) and YAML options. 5. Live-interpret training logs, validation curves, and evaluation metrics. 6. End with best-practice checklists for scaling from small experiments to full production runs. Attendees will leave able to replicate the full workflow—dataset → training → evaluation → deployment—entirely on their MacBooks.

- Event context: AI Tinkerers - Orange County Meetup- Wednesday June 4th 2025 — 2025-06-05 — Orange County
- Public talk page: https://orange-county.aitinkerers.org/talks/rsvp_S0EeHEPZUv4

### [Benchmarking 100 LLM Inference Engine Configurations](https://nyc.aitinkerers.org/talks/rsvp_XcK5R374aSw)

I will present stopwatch, an open source tool we built for benchmarking LLM inference engines (vLLM, SGLang, TRT-LLM) quickly -- both iteratively and massively in parallel. Time permitting, I will also show a few benchmark results that might be of interest.

- Event context: How It’s Made: Architecting Planning-Based AI Systems ft. AI21 Maestro — 2025-05-19 — New York City
- Public talk page: https://nyc.aitinkerers.org/talks/rsvp_XcK5R374aSw

### [The almighty function-caller](https://paris.aitinkerers.org/talks/rsvp_u7qReAn7jIw)

How would you like to give extensive tools memory to your edge agentic system, and optimize the resources it takes to run yet a high-performance set of agents ? We came up with a novel approach to function-calling at scale for smart companies and corporate-grade use-cases. The code is fully open-source.

- Event context: AI Tinkerers Paris: AI21 Labs Takeover on May 19th — 2025-05-19 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_u7qReAn7jIw

### [AI Patient Voice Simulator](https://seattle.aitinkerers.org/talks/rsvp_i7eeiSZT1Wo)

A voice to voice simulation of an Virtual AI patient using the Elevenlabs API. I will be demoing a project where I use the elevenlabs conversational agent to roleplay as a patient to practice diabetic patient consulting. This will be hosted on a nextjs front-end with a chatbot UI setup styled using shadcn components.

- Event context: AI Tinkerers Seattle - April Meetup — 2025-04-25 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_i7eeiSZT1Wo

### [OCR 2.0, Using Vision Language Models for instruction augmented OCRs](https://bengaluru.aitinkerers.org/talks/rsvp_iI8E7i1YB6E)

OCR is almost a solved problem but not really a generalizable problem. Although I am not an industry expert, but from talking to some expert I realized that classical (Object detection / text-recognition / rule based) OCRs normally parses the whole document and returns a very "less organized" results. Lot of manual post processing are involved on these kind of OCRs. The level of post processing / hardcoding logic for a family of documents rises with the increase in complexity of the document (example: if document contains combination of tables / images etc). With rise of LLMs and Vision Language Models (VLMs), the above problems can be solved and can be generalized over a range of documents. Last week, I started this new side project of mine, of building a more generalized OCR pipeline, where user will upload the document and also provide the expected output schema. The pipeline will do the OCR and would structure the result adhering to the uploaded schema. I tried over some range of documents (complex documents with tables, invoices and multi lingual documents). Now I do not want to make this another Open AI wrapper software, for several reasons: - For enterprise focussed documents, invoices might contain lot of PIIs and user would not be comfortable giving it to a third party client. - I did some bunch of experiments, and from there, I learned, I can achieve a very good pipelines with models < 7B parameters. I have been using ensembling approaches and it really works. For instance, there is a recent model called GOT-OCR 2.0, which gives awesome result but it is based on Qwen 0.5B. More open source models like InternLM, Qwen models, Llama 3.2 vision are also amazing and adheres with the schema. - It also gives me a full flexibility for furthur fine-tuning models on very complex documents where it fails to give results. Speaking of fine-tuning, I surely have faced challenges. For example, I have been fine-tuning GOT-OCR for a less known language, and I learned that it got overfitted on my training dataset and could not generalize over anything outside the training dataset distribution. Also for fine-tuning, generating data samples is also another challenge. In this talk I will be sharing my above learnings and my roadmap in more details and lead to an open discussion. ps: I will show the demo while doing the presentation

- Event context: AI Tinkerers Bengaluru - December - RSVP REQUIRED — 2024-12-05 — Bengaluru
- Public talk page: https://bengaluru.aitinkerers.org/talks/rsvp_iI8E7i1YB6E

## Related Technologies

- [llama](https://aitinkerers.org/technologies/llama) ([Markdown](https://aitinkerers.org/technologies/llama.md)) — 48 public demos
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 662 public demos
- [vLLM](https://aitinkerers.org/technologies/vllm) ([Markdown](https://aitinkerers.org/technologies/vllm.md)) — 33 public demos
- [LoRA](https://aitinkerers.org/technologies/lora) ([Markdown](https://aitinkerers.org/technologies/lora.md)) — 17 public demos
- [Ollama](https://aitinkerers.org/technologies/ollama) ([Markdown](https://aitinkerers.org/technologies/ollama.md)) — 77 public demos
- [Swift](https://aitinkerers.org/technologies/swift) ([Markdown](https://aitinkerers.org/technologies/swift.md)) — 17 public demos
- [unsloth](https://aitinkerers.org/technologies/unsloth) ([Markdown](https://aitinkerers.org/technologies/unsloth.md)) — 8 public demos
- [AutoWeeb](https://aitinkerers.org/technologies/autoweeb) ([Markdown](https://aitinkerers.org/technologies/autoweeb.md)) — 2 public demos
- [BAML](https://aitinkerers.org/technologies/baml) ([Markdown](https://aitinkerers.org/technologies/baml.md)) — 12 public demos
- [ChromaDB](https://aitinkerers.org/technologies/chromadb) ([Markdown](https://aitinkerers.org/technologies/chromadb.md)) — 10 public demos
- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 215 public demos
- [clustering](https://aitinkerers.org/technologies/clustering) ([Markdown](https://aitinkerers.org/technologies/clustering.md)) — 3 public demos
- [Emotional vectors](https://aitinkerers.org/technologies/emotional-vectors) ([Markdown](https://aitinkerers.org/technologies/emotional-vectors.md)) — 1 public demo
- [Gemini Nano](https://aitinkerers.org/technologies/gemini-nano) ([Markdown](https://aitinkerers.org/technologies/gemini-nano.md)) — 2 public demos
- [Go](https://aitinkerers.org/technologies/go) ([Markdown](https://aitinkerers.org/technologies/go.md)) — 25 public demos
- [GOT-OCR 2](https://aitinkerers.org/technologies/got-ocr-2) ([Markdown](https://aitinkerers.org/technologies/got-ocr-2.md)) — 1 public demo
- [Hugging Face](https://aitinkerers.org/technologies/hugging-face) ([Markdown](https://aitinkerers.org/technologies/hugging-face.md)) — 41 public demos
- [Hugging Face Hub](https://aitinkerers.org/technologies/hugging-face-hub) ([Markdown](https://aitinkerers.org/technologies/hugging-face-hub.md)) — 2 public demos