Junior Machine Learning Engineer

About Unbox Commerce

Unbox Commerce builds the operating system for social commerce. We help the world's most brand-sensitive companies scale creator-led commerce on platforms like TikTok Shop. Our platform turns social video into brand-specific intelligence: letting brands find the right creators, catch brand safety problems early, and figure out what's working.

We're a small team based in London, working with some of the world's largest brands. We've raised an investment round and we're growing. ML and AI are central to what we do.

How we work

We try to question how things should be done rather than copy other companies' defaults. We move quickly when there's a real customer problem to solve. There's no ego or politics on this team and we plan to keep it that way. We try to leave each week a bit better than the last.

The opportunity

We process millions of TikTok videos a day. Our pipelines pull videos, transcripts, and metadata from creators across the platform, run them through frontier model APIs (OpenAI, Anthropic, Google) for understanding and extraction, and turn the results into data products our customers rely on.

You'll write Python services that orchestrate LLM calls at scale, using Temporal for workflow orchestration. You'll extend our pipelines to cover new brands, sub-brands, and API fields, fix bugs, and ship features. You'll also pick up impromptu work: ad-hoc data and information requests, minor pipeline changes, and updates to our data warehouse models when business needs shift.

This is an applied engineering role. You'll use APIs for models. You will not train models from scratch. The hard problems here are reliability, throughput, cost, and getting useful structured output from frontier models.

You'll work alongside our senior ML engineer, picking up integration and enablement work from them so they can focus on harder research and design problems. You'll be talking to engineers, researchers, and people who don't write code, working across a mix of research code and production code. Communicating clearly about both matters.

• Location: London, hybrid: Mondays and Wednesdays in office
• Compensation: competitive salary
• Start: Flexible, ideally Summer 2026

What you'll work on

• Python services processing millions of videos a day, calling frontier model APIs for understanding, extraction, and classification
• Taking research code from senior engineers and integrating it into production pipelines
• Extending pipelines to cover new brands, sub-brands, and API fields
• Ad-hoc data requests and data warehouse model updates as reporting needs shift
• Writing documentation and explaining trade-offs to technical and non-technical audiences

Who we're looking for

This is an early-career role. We care more about how you think than how long you've been doing this.

• Comfortable writing Python: you've built things beyond coursework, whether that's at a previous job, an internship, a side project, or open source
• Comfortable with Git, CI, and SQL day to day, or able to get there fast
• Some experience with async programming in Python (or the willingness to learn quickly)
• Curious about LLMs and how to get useful structured output from them
• Strong written communication: you can explain a trade-off in a paragraph rather than a meeting
• Able to be in our London office on Mondays and Wednesdays

Nice to have

• A relevant degree (CS, ML, Maths, Engineering, or similar)
• Experience with Temporal or another workflow orchestration tool (Airflow, Prefect, Dagster)
• Experience with data warehouses
• Experience with big data (millions of records a day or more)
• Experience with frontier model APIs: prompting, prompt evaluation, fine-tuning
• Experience processing video or audio data
• Familiarity with ML tooling for production (MLflow or similar)
• Deeper cloud experience, particularly AWS

Why Unbox

• You'll ship ML code at real scale from day one: millions of videos, frontier model APIs, customers who notice when something breaks
• You'll work directly alongside our senior ML engineer, with a clear path to harder research and design problems as you grow
• Your code ships immediately, and you'll talk to the people using it
• Small team, no politics, no ego: easy to ask questions and learn fast
• You'll see how social commerce actually works, on the platforms where it's happening

How to apply

Submit your CV and a short note on what you'd want to work on via this Notion form.

We review applications on a rolling basis: if you're the right fit, expect to hear from us fast.

Unbox Commerce is an equal opportunity employer. We welcome applicants from all backgrounds.