Founding Machine Learning Engineer, Composo

Salary not provided

+ Substantial equity

Python

PyTorch

Junior and Mid level

London

4-5 days a week in office (Holborn, London)

LLM application evaluation for quality & accuracy

Be an early applicant

LLM application evaluation for quality & accuracy

1-20 employees

B2BArtificial IntelligenceDeep Tech

Be an early applicant

Salary not provided

+ Substantial equity

Python

PyTorch

Junior and Mid level

London

4-5 days a week in office (Holborn, London)

1-20 employees

B2BArtificial IntelligenceDeep Tech

Company mission

To help teams conduct precise, automated evaluation of complex LLM applications.

Job

Company

Role

Who you are

2+ years experience in machine learning, neural networks and GPU-accelerated training
Experience in Pytorch or similar modern machine learning frameworks (e.g. Huggingface, Llama.cpp, JAX etc)
Good Python programmer - comfortable with the level of software maturity required to organise data curation, augmentation and training workflows
You are results driven. A commercial and user-centric focus drives everything you do
You're capable of both running autonomously and collaborating closely
You're excited about collaborating with our London team in person

Desirable

Experience with NLP, transformer architectures and modern large language models (Llama, Gemma, etc)
Experience prompting, building and evaluating applications with frontier LLM APIs (OpenAI, Gemini, etc)
Experience working in early stage startups
Software development experience e.g. developing and serving REST APIs

What the job involves

Build upon our existing foundational R&D, leveraging the latest, leading research in LLM evaluation
Lead our ML development, training & scaling work
Work from a customer-centric perspective to iterate and improve models

Application process

1. Apply with your profile
2. 15 minute call with CTO
3. 60 min coding interview
4. 45 min chat with CEO
5. 45 min ML research interview

Salary benchmarks

Our take

The AI market is booming but LLMs still face huge reliability issues. Companies want to use these models but often struggle with inconsistent performance. Composo is stepping in to tackle this, helping businesses make sure their AI apps work as intended without the hassle of human testing or unreliable checklists.

It offers both a no-code platform and an API, making it easy for companies to assess their API apps. By using custom models that align with specific business criteria, such as medical chatbots detecting red-flag symptoms, Composo helps businesses ensure their AI is never off-point.

It's an exciting time for Composo as it has secured names like Accenture, Palantir, and McKinsey as customers and recently closed a pre-seed funding round. The funding will go towards supporting team expansion and ramping up R&D. With this solid foundation and clear vision, Composo is set to run the race in the AI reliability space.

Freddie

Company Specialist at Welcome to the Jungle

Company

Funding (1 round)

Dec 2024

$2m

SEED

Total funding: $2m

Company HQ

Holborn, London, UK

Leadership

Sebastian Fox

(Co-Founder & CEO)

Previously a Product Manager at QuantumBlack, AI by McKinsey, as well as an Associate and Engagement Manager at McKinsey & Company.

Luke Markham

(Co-Founder & CTO)

Previously a Machine Learning Engineer at Graphcore and a Machine Learning Engineer/Technical Lead at Tessella.

Share this job