# Optuna Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/optuna
> Markdown URL: https://aitinkerers.org/technologies/optuna.md
> Technology record last updated: 2026-02-25T10:18:22Z
> Generated: 2026-09-22T12:38:02Z

Optuna is the automatic hyperparameter optimization framework: it uses state-of-the-art algorithms to find optimal hyperparameter values fast.

Optuna is the automatic hyperparameter optimization framework you need for machine learning: it accelerates model tuning. The platform features a novel Define-by-Run API, ensuring high code modularity and dynamic search space construction. It employs state-of-the-art algorithms, like Tree-structured Parzen Estimator (TPE), to efficiently sample hyperparameters and prune unpromising trials, significantly outperforming traditional methods like GridSearch. For speed, Optuna supports easy, asynchronous distributed optimization, allowing studies to scale across tens or hundreds of workers with minimal code changes.

- Official technology site: https://optuna.org
- Public AI Tinkerers demos and talks: 3
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Building Chessboard Recognition App with my daughter](https://toronto.aitinkerers.org/talks/rsvp_zTY6p1BvBBQ)

Snap That Piece is the over-the-board chess companion working on the edge with no internet connection. Built as a dad companion to help keep up with his daughter's supersonic speed progression at chess, it's a side project born out of passion. I will be presenting the process from start to finish. From creating the data acquisition and labelling pipeline with a custom labeling software all the way up to training, tuning hyperparameters, and deploying an app as a wrapper to the CoreML model on Apple devices. I will also show how to transition from popular proprietary Ultralytics YOLO to the open-source Roboflow RF-DETR model family.

- Event context: AI Tinkerers Toronto - April 2026 - hosted by Shopify — 2026-04-29 — Toronto
- Public talk page: https://toronto.aitinkerers.org/talks/rsvp_zTY6p1BvBBQ

### [Evolutionary Agents with Algorthmic Prompt Optimization](https://amsterdam.aitinkerers.org/talks/rsvp_1SH7XblRcJA)

Demo of some recent research project at Comet on-top of DSPy's optimization algorthims as well as other hypertuners to improve agent prompts for a give model. I will show how this can be automated through MCPs to further enhance your given your agentic application.

- Event context: AI Tinkerers Amsterdam - May Edition — 2025-05-06 — Amsterdam
- Public talk page: https://amsterdam.aitinkerers.org/talks/rsvp_1SH7XblRcJA

### [AutoRAG](https://berlin.aitinkerers.org/talks/rsvp_4eSc1tq148s)

Automatic hyperparameter optimization of your retrieval augmented generation pipeline. Stop fiddling around manually with chunk size, top k and so on. Let the machine do the tuning of your retrieval settings. Inspired by hyperparameter optimization in conventional ML.

- Event context: AI Tinkerers Berlin - March 21 — 2024-03-21 — Berlin
- Public talk page: https://berlin.aitinkerers.org/talks/rsvp_4eSc1tq148s

## Related Technologies

- [AutoRAG](https://aitinkerers.org/technologies/autorag) ([Markdown](https://aitinkerers.org/technologies/autorag.md)) — 1 public demo
- [Ax](https://aitinkerers.org/technologies/ax) ([Markdown](https://aitinkerers.org/technologies/ax.md)) — 1 public demo
- [ClearML](https://aitinkerers.org/technologies/clearml) ([Markdown](https://aitinkerers.org/technologies/clearml.md)) — 1 public demo
- [Comet Opik](https://aitinkerers.org/technologies/comet-opik) ([Markdown](https://aitinkerers.org/technologies/comet-opik.md)) — 2 public demos
- [CoreML](https://aitinkerers.org/technologies/coreml) ([Markdown](https://aitinkerers.org/technologies/coreml.md)) — 2 public demos
- [DSPY](https://aitinkerers.org/technologies/dspy) ([Markdown](https://aitinkerers.org/technologies/dspy.md)) — 12 public demos
- [GPyOpt](https://aitinkerers.org/technologies/gpyopt) ([Markdown](https://aitinkerers.org/technologies/gpyopt.md)) — 1 public demo
- [Hyperopt](https://aitinkerers.org/technologies/hyperopt) ([Markdown](https://aitinkerers.org/technologies/hyperopt.md)) — 1 public demo
- [Keras Tuner](https://aitinkerers.org/technologies/keras-tuner) ([Markdown](https://aitinkerers.org/technologies/keras-tuner.md)) — 1 public demo
- [MobileOne](https://aitinkerers.org/technologies/mobileone) ([Markdown](https://aitinkerers.org/technologies/mobileone.md)) — 1 public demo
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 662 public demos
- [RAG](https://aitinkerers.org/technologies/rag) ([Markdown](https://aitinkerers.org/technologies/rag.md)) — 147 public demos
- [Ray Tune](https://aitinkerers.org/technologies/ray-tune) ([Markdown](https://aitinkerers.org/technologies/ray-tune.md)) — 2 public demos
- [RF-DETR](https://aitinkerers.org/technologies/rf-detr) ([Markdown](https://aitinkerers.org/technologies/rf-detr.md)) — 2 public demos
- [scikit-optimize](https://aitinkerers.org/technologies/scikit-optimize) ([Markdown](https://aitinkerers.org/technologies/scikit-optimize.md)) — 1 public demo
- [SMAC3](https://aitinkerers.org/technologies/smac3) ([Markdown](https://aitinkerers.org/technologies/smac3.md)) — 1 public demo