Technical Director, Generative AI at Jyá


Location: London, UK
Type: Full-Time
Category: Developer


Summary: Transformer-based models (e.g. Codex). £40k-50k plus equity. Not fully remote.

About Jyá

Jyá is a startup based in London working on software tools that are highly customisable without users needing to do any coding. Core use-cases centre around supporting highly flexible and highly expressive interactions with music data, enabling composers, producers and dj’s to create and perform music that they have imagined but until now have not had the tools to make a reality.

Who will I be working with?

My name is Lucas, I founded Jyá after I finished my PhD. If you join then to start with it will just be the two of us, but clearly the aim will be to expand the team as quickly as possible.

What kind of candidate is this role a good fit for?

This is a hands-on role across both software engineering and data science that would suite someone who has expertise in machine-learning in addition to a solid understanding of development across interactive apps and databases. While your focus will be the generative AI aspects of the work, you will need to be flexible and happy to work across all technical aspects as we develop prototypes and a proof-of-concept.

Music applications are complex use-cases and so getting the platform to the point where it is ready to support these will likely mean the platform will scale to other use-cases. It is important to understand that while the Technical Director position is suited to someone who is passionate about music and excited by the idea of working on music applications, there may be times when some or all of your focus will be on non-music applications.

From a technical point of view the main focus will be training / fine-tuning transformer neural networks (e.g. Codex) to return code that integrates automatically into client applications. More detail about this will be shared with candidates who progress through the hiring process.

Candidates will need a solid background in Python and JavaScript. An ‘ideal’ candidate would also have expertise in Swift, C++, graphics APIs, audio engines and graph databases. Clearly it is not realistic to expect someone to be an expert in all of these things, but this should give you a sense of how flexible you will need to be in terms of the breadth of the work.

What will I be doing?

You will be experimenting and building prototypes as we work towards a proof-of-concept, with your main focus being to support me as I work to secure additional funding to expand the team.

How much money will I make?

I am offering a salary of £40k-£50k plus equity. If you are a senior software engineer or data scientist then this is probably less than you currently earn, but you will have the benefits of working on a startup while still receiving a salary. There will be scope for the salary to increase when we secure additional funding.

How much equity will I get?

It is reasonable for candidates who are taking a larger salary hit to come on board to receive a larger stake, so this will be something to discuss with candidates on a case-by-case basis.

What are you looking for / are there any pre-requisites?

The focus of the work at Jyá is to develop technologies and products for the direct expression of creative thought. I am looking for someone who shows they grasp what this means and are excited by the possibility of working on this mission with me. Beyond that, I am essentially looking for someone who convinces me they will be capable of working with me to deliver prototypes.

If you have read this far you will have realised that there is quite a demanding balance of skills required for this role. While candidates will need to show clear expertise in machine learning, the day-to-day work will not be limited to this, and so whatever your background, you will need to demonstrate somehow that you have the balance of technical / creative / practical skills required for success.

I am not specifying pre-requisites in terms of academic background, however it is unlikely successful candidates will not have completed advanced technical degrees. I will, for example, ask things like: ‘talk me through the process when you created your own implementation from a paper, when the authors’ approach was compelling but they didn’t provide any code’.

Demonstrating practical experience with building and training transformer-based models will be a clear advantage and candidates will need to evidence a strong grasp of the mathematical foundations required for the role (e.g. linear algebra, calculus, statistics).

Where will I work?

While this role will be mainly remote to start with it is not suitable for candidates who are looking for 100% remote. Travel expenses won’t be covered and you will need to be in London some of the time, so if you are not London-based this is something to consider.

If you love the sound of this opportunity but you are not based in the UK or are looking for 100% remote then feel free to reach out so you can be notified of other opportunities in the future.

How To Apply:

For the candidate who is successful the process will be:

  • Email your CV and cover letter (PDF) to the main inbox (, including any links you want to share in the cover letter document.
  • Provide written answers to questions that are sent to you.
  • Attend a video call intro.
  • Provide written answers to additional questions.
  • Attend an in-person interview.
  • Offer.

No fixed end date for applications. The application process will close when the right candidate is found.