AI / ML Engineer
Build AI systems that reach production with measured accuracy — document extraction, retrieval, prediction and agents with real boundaries.
Build AI systems that reach production with measured accuracy — document extraction, retrieval, prediction and agents with real boundaries. This is a remote full-time role at Devinity Solutions LLC, a software engineering firm of 60 people founded in 2016.
About the role
We build AI into software that runs every day: freight document parsing, property valuation modelling, autonomous vendor ordering. All of it is measured against an evaluation set before it ships.
If you have shipped an AI feature and then had to defend its accuracy to someone who depended on it, this role will be familiar.
What you will do
- Define correctness with clients and build evaluation sets from real data
- Build retrieval, extraction and prediction systems in Python and TypeScript
- Establish accuracy baselines and score every change against them
- Design confidence thresholds and human-in-the-loop review
- Instrument accuracy, latency and cost in production
What we are looking for
- Experience taking an AI system from prototype to production
- Strong Python, and comfort with TypeScript
- Practical understanding of evaluation: test sets, baselines, regression
- Experience with at least one major model provider API
- Ability to tell a client when AI is the wrong tool for their problem
Useful, not required
- Self-hosted open model deployment
- Vector search and retrieval system design
- Document processing at volume
How to apply
Email us with what you have built and why it was difficult. A CV is fine, but the thing we read first is the description of a hard problem you solved.
Apply for this role