not-a-cat

Tufalabs | Zurich, ONSITE | Full-time | 100-200k CHF | https://tufalabs.ai/

(I want to preface that we don't have a ton of budget to expand our team right now and we want to keep the team small, but I also don't want to miss someone who would be a great fit)

We are a self-funded independent AI research lab in Zurich. Currently we explore ARC-AGI3, LLM/RL reasoning, and quant. We recently started a new LLM post-training and pre-training effort, led by some ex-engineers from the swissAI Apertus team. We also like to do some Kaggle on the side, our team includes the top #3 and #12 globally.

Our overall roadmap is to catch up with LLM capabilities, own the full stack, and use it to do our own independent research, pushing efficiency and looking for new paradigms. We expect our new cluster (2x NVL GB300) to be up by the end of the month.

Long term we want to switch to AI alignment/safety research as I believe this is all that matters when AI capability is high enough.

We look for people who are driven, who believe it's possible to make significant contributions to both AI capabilities and safety, and with previous experience pre-training and post-training LLMs and related infra.

If interested, email me at benjamin at tufalabs ai.

This wouldn't happen if employees rejected cash-based compensation and decided to be founders themselves. Most employees trade risk for higher cash comp, and end up with less upside. This issue is mostly settled by the employment market

Tufalabs | Senior research/engineers | Full Time | Zurich, Onsite | 100-200k CHF | https://tufalabs.ai/ Our current focus is arc-agi 3. We recently hired the top #2 kaggler to join that effort. We are looking for people deeply passionate about AI research. I am steering our work culture towards "live-to-work": If you wouldn't be doing AI research working somewhere else, then it is not a good fit. If you don't find yourself naturally reading about AI papers or doing random AI projects in your free time, then it is not a good fit. We have high expectations and a generous equity structure to match.

Tufalabs is an independent AI research lab in Zurich, focused on high-impact academic-style AI research and large-scale LLM development. We care about compute (currently, 6× DGX B200, 2× NVL72 GB300 incoming). Our mission is to push the frontier of AI science while open-sourcing results and building the engineering muscle to pre- and post-train cutting-edge language models. The ultimate goal is automating AI research, with a strong focus on alignment and safety in the limit.

We can only sponsor visas for people with EU nationality. The position is onsite, fulltime. We are self-funded, from what i call "self-propelled machine learning": I applied ML to cool domains and made returns that way. I can give more details about this privately. If interested, email me at benjamin at tufalabs.ai

Tufalabs | COO & Senior research/engineers | Full Time | Zurich, Onsite | 100-130k CHF | https://tufalabs.ai/

Tufalabs is an independent AI research lab in Zurich, focused on high-impact academic-style AI research and large-scale LLM development. We care about compute (currently, 6× DGX B200, 2× NVL72 GB200 incoming). Our mission is to push the frontier of AI science while open-sourcing results and building the engineering muscle to pre- and post-train cutting-edge language models. The ultimate goal is automating AI research, while keeping in mind AI alignment and safety. The ultimate goal is to make sure AI goes well for humanity.

I am looking for a COO to help scale from ~10 to ~15 people.

Ideally you care about AI research, you are excited about the idea of helping us move faster, you like to keep red-tape low, you are a technical person who can create automated processes, and have experience owning and running operations at a startup.

We are also open to senior engineers and researchers, comp negotiable

If interested, email me at benjamin at tufalabs.ai

I founded Tufalabs about a year ago https://tufalabs.ai/ On the side, I'm making a snake game with codex-cli, and then making a ppo agent to learn to play it. Codex-cli can mostly oneshot the ppo implementation, which I find extremely impressive. Getting codex to tweak the ppo algorithm resulting in a RL agent fully solving the snake and visualizing it is very satisfying.

Tufalabs | COO & Senior research/engineers | Full Time | Zurich, Onsite | 100-130k CHF | https://tufalabs.ai/

Tufalabs is an independent AI research lab in Zurich, focused on high-impact academic-style AI research and large-scale LLM development. We care about compute (currently, 6× DGX B200, 2× NVL72 GB200 incoming). Our mission is to push the frontier of AI science while open-sourcing results and building the engineering muscle to pre- and post-train cutting-edge language models. The ultimate goal is automating AI research, while keeping in mind AI alignment and safety. The ultimate goal is to make sure AI goes well for humanity.

I am looking for a COO to help scale from ~10 to ~15 people.

Ideally you care about AI research, you are excited about the idea of helping us move faster, you like to keep red-tape low, you are a technical person who can create automated processes, and have experience owning and running operations at a startup.

We are also open to senior researchers, comp negotiable

If interested, email me at benjamin at tufalabs.ai

Tufalabs | COO & Senior research/engineers | Full Time | Zurich, Onsite | 100-130k CHF | https://tufalabs.ai/

Tufalabs is an independent AI research lab in Zurich, focused on high-impact academic-style AI research and large-scale LLM development. We care about compute (currently, 6× DGX B200, 2× NVL72 GB200 incoming). Our mission is to push the frontier of AI science while open-sourcing results and building the engineering muscle to pre- and post-train cutting-edge language models. The ultimate goal is automating AI research, with a strong focus on alignment and safety.

I am looking for a COO to help scale from ~10 to ~15 people.

Ideally you care about AI research, you are excited about the idea of helping us move faster, you like to keep red-tape low, you are a technical person who can create automated processes, and have experience owning and running operations at a startup.

We are also open to senior researchers, comp negotiable

If interested, email me at benjamin at tufalabs.ai

Tufalabs | AI researchers and engineers | Zurich | Full time | Onsite | $165k | https://tufalabs.ai/

We are a small, independent research group working on fundamental AI research.

Since around 2008, I always believed that AI research will be the most impactful technology to work on, both from an intellectual curiosity and technological capability perspective. After a few failed attempts to get into ML/AI research, I decided to apply it to quant finance, and then start an AI research lab from it.

This is how Tufalabs was funded about a year ago. We started working on ARC (https://arcprize.org/arc-agi), and since o3 we focus on LLM-based reasoning, self-improving loops and scaling trends. You can see some of our research there: https://tufalabs.ai/blog/_site/

We do not have too many spots left, but I want to make sure we don't miss-out on people who would be great fits. We look for obsessive people who share the same research interests. To apply, email me at not_a_cat at fastmail com.

Tufalabs | Research engineer | Zurich | Full Time | On-site

We are a small, independent research group working on fundamental AI research.

The plan is to first solve the ARC benchmark (https://arcprize.org/), then use that to secure more funding, and move on to more difficult benchmarks. The research topics we want to investigate next are test-time compute, active inference, LLM-based reasoning and system2 reasoning. The ultimate goal is to do cool AI research leading up to AGI.

The lab has acquired the MindsAI team recently. MindsAI is the best team in the competition so far, pioneering techniques such as fine-tuning models on synthetic data, test-time-training, and test-time-finetuning.

We are looking for people who are driven and intrinsically motivated to work on both ARC and AI/ML research. The role would include working on the main system developed by MindsAI to accelerate progress.

Compensation: $100-135k/year + generous equity plan

More info: https://tufalabs.ai/

To apply: send me an email at [email protected]

Hope you get it sorted. I have seen relatives struggle through it and it is a challenge

Unnamed AI research lab | ML research engineer | Zurich, Switzerland | Full-time | Onsite | 100-200k CHF I am starting a research lab funded by a hedge fund that I created 5 years ago. The hedge is doing well, and applies ML to quant finance. The research and trading activities should reinforce each other over time: the research publishes research and fosters in-house ML expertise. The trading applies ML to quant finance and brings funding. The research entity is incorporated, funded, has an office and 0 employees currently.

My vision for the research lab is the following:

  - A bunch of phds in a room: create a good working environment, optimize for research output & runway.
  - Take the "bitter lesson" seriously: most of the successes in modern machine learning are due to deep learning and the ability to leverage vasts amounts of compute and data.

I want to study how DL/LLM models reason, using ARC as a benchmark (https://arcprize.org/, https://arxiv.org/abs/1911.01547). Techniques I want to investigate include test time compute, active inference, meta-learning, self-play, or any technique that would be considered "system 2".

With that in mind, today I'm looking for the following role:

  - ML research engineer

Compensation will include flexible options scheme based on results. Interest and knowhow in the following is a plus: Deep learning, transformer models, reasoning, world models, learning-to-learn, mechinterp, bayesian methods, probabilistic programming.

I will mostly consider candidates having an EU passport for visa reasons.

To apply, email me at not_a_cat at fastmail dot com.

Unnamed AI research lab | ML research engineer | Zurich, Switzerland | Full-time | Onsite | 100-200k CHF

I am starting a research lab funded by a hedge fund that I created 5 years ago. The hedge is doing well, and applies ML to quant finance. The research and trading activities should reinforce each other over time: the research publishes research and fosters in-house ML expertise. The trading applies it to quant finance and brings funding. The research entity is incorporated, funded, has an office and 0 employees currently.

My vision for the research lab is the following:

  - A bunch of phds in a room: create a good working environment, optimize for research output & runway.
  - Take the "bitter lesson" seriously: most of the successes in modern machine learning are due to deep learning and the ability to leverage vasts amounts of compute and data.

So we make it a top priority to develop the infrastructure and expertise to be good at DL/LLMs/transformers. It doesn't mean that we don't explore new ideas. We want to learn to leverage large amounts of compute and be smart/efficient about it. If the trading works well, we can compete on an equal footing with the big labs over time in terms of amount spent in compute.

For the research itself, I want to start with the ARC challenge (https://arcprize.org/, https://arxiv.org/abs/1911.01547)

With that in mind, today I'm looking for the following role:

  - ML research engineer

Compensation will include flexible options scheme based on results.

Interest and knowhow in the following is a plus: Deep learning, transformer models, reasoning, world models, learning-to-learn, mechinterp, bayesian methods, probabilistic programming I'm also looking for evidence of past exceptional achievements.

I will mostly consider candidates having an EU passport for visa reasons.

To apply, email me at not_a_cat at fastmail dot com.

Tufa.ai | SRE, SWE, operations manager | Zug, Switzerland | Full-time | Onsite | EU much preferred | 130k-260k CHF

5 years ago, I started a side project to apply ML to crypto trading to finance my own AI research. It has now evolved into a hedge fund with 10 people, 100M+ AUM (mostly prop capital) where we apply ML techniques to trade both crypto and US equities markets.

In addition to creating a sister entity to do fundamental AI research, we want to grow the quant team a bit more. We are looking for the following roles:

- SWE+SRE for infrastructure: exchange connectivity & internal trading software | C++, typescript, some python, AWS

- SWE+quant | Python, PyTorch

- Operations manager | Experience with prop or hedge funds, working with lawyers & accountants, reporting, working with brokers/exchanges in both traditional finance and crypto. Familiarity with scripting and automating processes is a plus.

Our compensation package usually includes the ability to re-invest bonuses in the fund, a bonus scheme (using 30% of profits after costs), and some options.

For all roles, we will hire opportunistically and only consider very strong candidates. Interest for trading/quant finance and/or ML is a big plus.

To apply, email me at not_a_cat at fastmail dot com.

Discussion on twitter: https://twitter.com/shr_id/status/1602688734990053376

Research & prop trading firm | Zug/Zurich | Full time | On site (Visa sponsor)

We apply ML to algotrading. We plan to use the proceeds to fund a fundamental AI research lab. Our primary goal is making progress toward AGI. Over the last two years, we turned 5 digits into 9 digits, trading a volume of around 1-4B per day. We operate as a prop trading firm of 7 people, we are restructuring our operations and moving to Zug. The short term priorities are growing the trading team and starting the research operation.

We look for the following roles:

- Engineer (trading side): Design, implement and maintain the next version of our trading engine. Experience with market micro structure, writing robust code and dealing unreliable APIs required.

- Engineer (ML side): Automate and maintain our current trading models. Improve retraining, automated deployment and capital allocation systems.

Compensation: 200-300k, bonus scheme with reinvestment in the fund, 1-3% options with 4y vesting. To apply, email me at: not underscore a underscore cat at fastmail dot com.

Keywords: Trading side: cryptocurrency bitcoin ethereum machine learning deep learning algotrading market making high frequency trading limit orderbook liquidity provision market microstructure binance okex ftx python pytorch pandas numpy ssh aws typescript postgresql parquet pycharms numba rust backtest uniswap-v3 automated market maker liquidity pool arbitrage yield farming funding arbitrage prop trading hedge fund Research side: probabilistic programming simulation generative models learning-to-learn inference modelling data augmentation causality self-play GANs transformers bayesian-inference

AlgoTrading firm | Lead Software engineer | Onsite, Zurich area | Full-time

We apply ML to trading in order to fund fundamental ML/AI research over the long term.

We turned 40k into 9 digits over the last 2 years, trading 1 to 4% of the crypto markets (2-6B per day), and grew a team of 7 people. We are now relocating to Zurich, establishing a new corporate structure and growing the team.

Compensation for everyone is still under debate, but will involve salary + options + exposure in the fund. We use the following: python, pytorch, node, javascript, typescript, websockets, postgres, parquet, aws, cuda, scp, ssh, lambda cloud. We are experimenting with: c++, go, rust, self-play, MCTS, probabilistic programming, deep RL.

We are looking for strong engineers, primarily people who can take ownership of our trading engine. To apply send an email to base 64 decode of bm90X2FfY2F0QGZhc3RtYWlsLmNvbQ==

AlgoTrading firm | Remote (EU Timezones prefered) | Full-Time | COO, Technical lead

One and a half year ago, my deep learning crypto algotrading project turned into a prop trading firm. The long term goal since the beginning is to use the profits to fund ML/AI research. Our daily volume is around 2-5B per day (1-3% of the overall markets), profits last year was 9 digits, sharpe ratio between 7 and 12, max dd of 10%. We started as 2 founders and are now a team of 5.

We are looking for the following roles:

- COO: Talk to lawyers and accountants to help us setup the business correctly. Also manage assets custody, compensation, incentives, payroll, bills, taxes...

- Tech lead: Take ownership of our trading engine, understand in detail how it works and make decisions about a refactor or rewrite. Make it robust so that we do not lose 1M when the FTX api goes down. We are open to a rewrite in python, go, rust or C++.

Compensation for either role: 0.5 to 1.5% equity, 0.5 to 1% flat bonus on total profits, 100-150k salary. We also have other bonuses in place for very strong performers.

Tech Stack: ML: python/pytorch/jupyterlab. Engine: typescript/node/postgres/aws

In interested email me at not_a_cat at fastmail com

I spent 1 year learning about AI/ML with books and moocs. Then 1 year building a prototype. Then 1 year + 2 months as a team.

Agreed. Good question. Regarding our current approach, my hope is that someday our alpha goes away entirely and we can publish our strategy in detail. But going forward, the plan is to have a trading division that keeps research secret, and a fundamental AI research division (unrelated to trading) that publishes publicly.

We have a hard time believing it ourselves too. I did leave out that around 25M of it is profits reinvested over time and it did do well because we are in a bull market. Binance forces you to hold 11k BNB to be vip9 to get the best fees, we started buying it at around $15, it's worth now around $300 per BNB, and we also hold other cryptos as a side investment. The rest is algotrading profits.

The initial investment was 40k, no other funding rounds or money laundering.

I do not see your email in your profile. (Not OP, but really interested to talk about starting a research institute)

Most of it comes down do conviction and sticking to my not-so-promissing setup for 1 year, while every single of my peers were either thinking WTF are you doing, or telling me why my thing would fail. The rest is luck where we could scale way past what we though was possible.

I stayed focused by creating a penalty system where if I worked for less than 7 hours that day, I'd burn 1$ for every missing hour. I still maintain that spreadsheet with the self-debt over time (that I have yet to pay, but I will by burning crypto). Also my screen saver message is "get back to work"

I think it's both easy and hard, depending how you look at it. Hard because it's a tough ML problem where you have to combat overfitting and the messiness of the real world; easy because crypto is clearly one of the easiest markets to trade and there are opportunities all over the place (compared to equities for example). On top of my head: stat arb, yield arb, exchange arb, funding arb, market making / liquidity providing (CEX/DEX), doing/preventing sandwich attacks, good old yield farming...

There are idealists out there who get all worked up with 'watch the beauty of it'. Yeah in practice it will devolve into blood, sweet and tears, but it matters that you have the right spirit and can appreciate the fleeting moments of beauty

Hi. I went through this journey myself. I have a CS master degree, and always wanted to transition into AI research (focus on probabilistic programming, deep learning and self-play, but also simulation, learning-to-learn, bayesian inference, genetic algorithms). The initial plan was to self-learn machine learning, deep learning and statistics. I did that for 1 year, with the hope of joining on of the big AI labs afterwards. As it's very competitive, I couldn't join one of those labs, and wasn't willing to compromise on that. So I turned to plan B: salvage my ML skills by applying them to crypto trading, and if it works, then I'll have funding to start my own lab.

Plan B worked: in 1 year and 2 months, we turned 40k of our own funds into 100M. Today we are a team of 5 and we do $5B of daily volume, around 3% of the global crypto markets. Now the challenge is to start the research division, operate the trading and research in tandem, and create the right incentives: the research converts money into ideas, the trading ideas into money. The research would have two sub-divisions: fundamental and applied to trading.

My motivation for the research is mostly curiosity: I want to de-mystify intelligence (natural or artificial) and reverse-engineer it. When you talk about your lab, I resonate a lot with "watch[ing] the beauty of it". The way I would describe my research agenda is the following: put a bunch of phds in a room, see what happens.

If that sounds interesting, email me at [redacted]

Market making firm | On-site | Europe/US | CFO | Full-time | $200-500k, equity, stake in trading strategies

I am the founder of a trading firm with a team of 3 people. We are applying quantitative methods to market making in many trading instruments in the crypto space. I can disclose more details (returns/performance, location... in a private chat)

Our legal and accounting situation is very complex. We re looking for someone who can take over all fiscal/legal/corporate matters, help us design the correct incentives for everyone (equity/bonuses) and restructure the business. We want someone willing to relocate and work on-site with us.

What we look for: high implication, integrity, perseverance, ambition, appetite for high-risk/high-reward situations. Bonus points if you speak many languages (especially french, also german), and have experience with tax optimisation and complex corporate structures.

If interested, email me at [email protected]