Machine Learning Engineer
BACKGROUND
We’re building bloop, an open-source AI code-search tool that helps engineers find, review and understand code. bloop combines LLMs with semantic code search to answer natural language questions about private codebases, while also supporting super-fast regex search and precise code navigation.
Since launching, we’ve introduced thousands of engineers to bloop, reached over 3k stars on GitHub, and raised our seed funding round from leading investors including Y Combinator, Khosla Ventures, Sands Capital and LocalGlobe.
THE ROLE
We’re using LLMs to build a developer assistant that can interact with and answer questions about private codebases.
As a Machine Learning Engineer you’ll work right at the heart of this. Whether that’s building a vector-based semantic search engine to find relevant code to include in an LLM prompt, distilling knowledge into low-latency smaller models, or prototyping new ways for LLMs to interact with tools.
You'll join a small but dedicated team of engineers who want to build the future of software engineering. You'll get to show off too, as bloop is completely open-source.
At bloop, you will:
- Design and prototype pipelines that compose LLM calls
- Implement training and evaluation pipelines for model fine-tuning
- Optimise models to run on consumer hardware
- Stay up to date with LLM/NLP research, bringing the latest advances into production
Requirements
- Practical ML experience, whether at a large company or a startup
- Strong grasp of the modern LLM landscape (models, prompting, fine-tuning)
- Interest and understanding of vector search and contrastive learning
- Experience training models in the cloud (we use AWS)
- Strong ability to communicate technical concepts
- (Bonus) Experience with Rust
- (Bonus) Experience deploying ML models using Kubernetes
- (Bonus) Interest in programming language implementation (parsers, type-systems, compilers)
ABOUT US
bloop is a Y-Combinator backed startup building the next big developer tools company, like GitHub or StackOverflow. We use machine learning to automate menial tasks in the software development lifecycle, increasing the productivity of software developers.