In regulated financial advice, a wrong answer is a compliance breach. Figen AI builds AI systems that wealth professionals use in front of their clients and can defend in front of their regulator. We are hiring senior engineers in India to work on the core of that system, remotely, with our team in Paris.

Who we are

Figen AI is a Paris-based company building AI agents for wealth management professionals: independent financial advisers, family offices, private banks and insurers. Our platform, Figen OS, prepares client meetings, writes meeting notes and regulatory documents, and ties every recommendation to the client's profile and goals under European rules (MiFID II, IDD). In 2026, Figen AI won first prize for the most innovative company in the Palmarès Gestion de Fortune, voted by 4,000 French wealth professionals.

A mission-driven company

Figen AI has been a société à mission since it was founded in January 2025. Under French law, this status writes a social and environmental purpose into the company's bylaws. Ours: to reveal everyone's potential through accessible, personalised and innovative solutions. We have committed to three goals: AI tools that help people reconcile their convictions, goals and financial constraints; sustainable and inclusive finance that more people can access; and an environmental footprint for our infrastructure that we measure and reduce. Once a year, the team also runs a volunteer project for a charity (the Secours populaire français in 2025, Moi & Mes Enfants in 2026).

What we believe

AI should give advisers back the hours they spend on paperwork, so they can spend them with clients. We work for one profession only, and we know it in detail. Client data is hosted in France, and every action an agent takes has to be explainable to the client and to the regulator.

Our values

Autonomy. You get real problems from the start and organise your own work to solve them, wherever you work from. Autonomy goes hand in hand with solidarity: a colleague who is stuck becomes the whole team's concern.
Ownership. You carry your work through to the result and keep your commitments, to clients and to the team. The company holds itself to the same rule: an independent third party verifies its commitments as a société à mission.
Excellence. Our clients are wealth experts and expect expertise equal to theirs, in law, tax, regulation and engineering. Everyone applies that standard to their own work first and keeps their expertise sharp through reading, research and training.

Why the work is hard

Specialised models. Figen AI holds a GENCI compute allocation on Adastra, a French national supercomputer operated by CINES, to train models for financial advisory tasks.
Agents inside regulated workflows. Each step is logged and checked against the rules that apply to the client.
Evaluation. We measure quality, regulatory compliance and latency on real advisory tasks.
Production. Professionals rely on the system for client work every day, so reliability and running cost matter as much as model quality.

What you will do

Design, build and ship core parts of our AI platform, from prototype to production.
Train, fine-tune and evaluate models on domain tasks, on large GPU clusters.
Build agent systems that stay reliable under real load: tool use, planning, guardrails.
Build the datasets and evaluation pipelines that tell us, with evidence, whether a change is an improvement.
Make systems faster, cheaper to run and easier to observe.
Work directly with the engineering team in Paris. Your work reaches clients within weeks.

You might come from

We are open on background. Strong candidates usually come from one of these paths:

Machine learning or research. You have trained or fine-tuned models and you trust careful evaluations more than leaderboards.
Backend and agents. You have built distributed systems in production and now build LLM applications that hold up with real users.
Data and infrastructure. You have run GPU workloads, data pipelines or large-scale compute, and other engineers move faster because of your work.

What we expect

Five years or more of building software in production, or equivalent depth in research.
Strong Python and solid fundamentals: testing, code review, system design.
Hands-on experience with machine learning or LLM-based systems.
You can take a vague problem to a shipped result without close supervision.
Clear written English and the habit of working asynchronously with a team in Europe.

Nice to have: distributed training, GPU programming, reinforcement learning, retrieval, publications or open-source work, experience in finance or another regulated industry. French is not required.

Our pace

The pace at Figen AI is demanding: releases follow each other closely and priorities move quickly. This role suits people who like working at that intensity.

What we offer

A full-time remote contract from anywhere in India, as an independent contractor or through an employer of record.
Compensation: competitive, based on profile.
Access to large-scale compute when the work justifies it.
Short decision cycles and direct contact with the founders.
Three to four hours of daily overlap with Paris working hours.

Hiring process

Apply with your CV, LinkedIn profile or GitHub. No cover letter.
A 30-minute call.
A technical deep dive on something you have built. No puzzle questions.
A time-boxed exercise close to our real work.
A conversation with the founders.

Shortlisted candidates hear from us within two weeks.

Figen AI hires on skills and evidence. We welcome applications from everyone, including people with disabilities.

Follow Figen AI on LinkedIn for engineering updates and future openings.

Senior Software Engineer, Applied AI & Distributed Systems