freeCodeCamp x Encode | AI - Jev Workshop

# Hands-on workshop on Jev: a "System One" decision model to give your AI agents a fast, cheap brain for making decisions safely, locally, and creatively!

Registration to this event is only valid through the Luma platform: https://luma.com/w6i73bgm

Hello code campers!
In today's world where incidents like OpenAI hacking HuggingFace and even government departments have been the talk of the town, the next question that comes to mind is, how can we make those AI systems we are building safe? There's the left side of the spectrum, don't use them at all or use deterministic algos, and there is the right side of the spectrum, use more AIs to guard those swarms. Somewhere in the middle there's this idea of limiting the number of actions an LLM can perform to a limited set (carefully crafted to avoid every unwanted scenario). This also comes with ~400x (e.g: Jev) improvement in speed, latency and cost compared to the frontier models offered by OpenAI/Anthropic.

That's the story behind Jev ( from TypeSafe AI ), released 15 Sep 2026 out of stealth, by an ex-OpenAI dev, which definitely needs a closer look. So let's see and if it is worthy of the hype!

Instead of generating text token by token like an LLM, Jev takes a "state" plus a set of multiple-choice questions defined in code, and returns typed decisions all in parallel. This has many practical use-cases in ticket routing, phishing detection, lead scoring, browsing, final boss in a game you're developing, LLM-as-a-judge eval, security guardrails, etc..

Jev is closed source, so in line with our local-first tradition we'll also look at the open-source alternatives popping up. Ollama now supports Jev-style decision models like nimble, `tev1`, Laya, the Ollaya runtime, and other projects like Cloudflare's Clef, Kev and Von. We'll run them locally and compare against the hosted API (with the downside of having no reasoning traces). Some community benchmarks say small local LLMs (like Qwen) are closer to Jev than the hype suggests.

We'll go through the setup live, calling Jev via its API/OpenRouter, then running the open-source models on your own hardware.

This will be our 9th event and we hope we will be able to squeeze another one b4 the new year!

You can install Ollama ahead of the event to save time and also, feel free to check the slides from all the prev events: andysign.github.io/ai-london-01 , andysign.github.io/ai-london-02 , andysign.github.io/ai-london-03 , andysign.github.io/ai-london-04 , andysign.github.io/ai-london-05 , andysign.github.io/ai-london-06 , andysign.github.io/ai-london-07 , andysign.github.io/ai-london-08

This workshop is open to all levels, from beginners to experts.
Bring your laptop (make sure it's fully charged as local models tend to be power-hungry), and don't forget to bring your coding gear (hoodie).

***

We are also delighted to announce that we are going to be accompanied by a guest speaker: Lesia Tkacz. She will give us a flash talk on how we can apply today's workshop creatively:

Lesia: We’ve all generated sensible texts with language models, but if you're like me, you’ve probably messed around with plenty of funny and weird outputs too. Here is your chance to push further and reframe your model experimentation as a circuit-bending hack project.

This talk is a flash ‘how-to’ and pep talk to challenge you to go to the very fringes of creative possibilities with bleeding-edge model technology and old-school ‘Lo-Fi AI’ methods to create your own generative literature pieces.

I’ll take you on a fast-paced run through the strange world of Generative Literature, arm you with the top how-to tips based on my own research, and point the way to how you might apply all this to the latest AI models.

Finally, the end of the talk will reveal my true agenda; to build you up towards submitting to the National Novel Generation Month (NaNoGenMo) challenge this November. This has been running on GitHub since 2013, and is a platform where beginners and seasoned programmers alike have shared both lofty and irreverent creative text generation projects which span 50,000 words or more.

Lesia Tkacz has a background in sociotechnical AI topics and NLP, and focused her PhD research on understanding and developing human evaluations of creative generated text in the ‘pre-ChatGPT’ era. She’s also unusually obsessed with the history of AI & computing, the occult, and has several unfinished art pieces in the works. She currently labours as a Technical AI consultant.

***

Agenda for the evening (Subject to change):
18:00 - 18:30 Arrival and networking
Mingle with fellow developers over refreshments.

18:30 - 18:45 Introductions
Intro and a bit of audience engagement

18:45 - 19:15 What is a decision model?
Jev explained: System One vs System Two, benchmarks, pricing, and where it fits in an agent stack (and where it doesn't).

19:15 - 20:05 Workshop on Jev & open-source alternatives
Hands-on technical workshop (if you are non-technical it's OK, tag along). We'll call Jev live via the API, then run nimble/tev1/Laya locally and compare decisions side by side.

​20:05 - 20:15 Announcement (PSA) from one of our volunteers - Vlad S. is going to talk about his new project ( Vlad is a software engineer with 20+ y of exp delivering high-quality software for businesses, he developed a specific method for integrating local AI models into professional workflows, prioritising code quality, efficiency, & control ).

20:15 - 20:35 Guest speaker - Generative Literature and creative uses of Jev with Lesia Tkacz

20:35 - 20:50 Q&A and networking
Questions from the workshop, open mingling.

20:50 - 21:00 Closing remarks

END: Carry on the fun at a nearby pub which is right @ the corner.

***

## Event Venue

As always we'd like to extend a heartfelt thank you to Encode Club.
Encode Club has spent over five years fostering a global community of developers, researchers, and founders in Web3, AI, and emerging technologies. The Encode Hub brings everything together under one roof, providing a dedicated environment for collaboration and innovation through meetups, co-working, and in-person events.