The retrieval engine built for agents
Hornet is an ambitious startup building the next generation of retrieval infrastructure. Our engine powers agentic systems with an API surface that agents can configure, deploy, feed, query, and optimize autonomously. Building on battle-tested technology, serving billions of users, petabytes of data, and millions of requests per second, we're rethinking retrieval for the agentic era.
We have an office in the Bay Area to tap into market opportunities in the region, while the engineering team is based in Trondheim. We serve enterprise clients globally, enabling them to get quality work done with agents.
The role
We are looking for a Senior Research Engineer who lives at the frontier of Information Retrieval (IR) and LLMs.
At Hornet, we believe that agents are the new users of search. This role isn't about optimizing click-through rates for humans; it's about optimizing retrieval for agents. You will lead the research into how agents interact with knowledge, defining what "relevance" means when the consumer is an agent rather than a person.
You will be the scientific heartbeat of our retrieval engine. You will be responsible for defining what "quality" looks like for agentic retrieval, running rigorous experiments to prove it, and ensuring our customers have the most effective retrieval and ranking strategies for their specific agentic workflows.
You will bridge the gap between theoretical research and production-grade engineering, turning experimental wins into core product features.
What you'll do
- Design and execute retrieval and ranking experiments on public benchmarks such as BEIR, BrowserComp-Plus, SimpleQA, and others, as well as proprietary customer datasets to ensure Hornet remains the gold standard for agentic retrieval
- Act as a technical advisor to our enterprise customers and solution services organization, helping them navigate the complex landscape of embedding models, re-rankers, and hybrid search strategies
- Translate research insights and experimental results into actionable feedback for the core engineering team to improve the engine's performance
- Produce high-quality technical content, including whitepapers, blog posts, and documentation
What we're looking for
- 5+ years of industry experience in Machine Learning, with a specific focus on Natural Language Processing (NLP) or Information Retrieval (IR)
- Strong understanding of dense and sparse retrieval, cross-encoders, embedding models, hybrid search, and evaluation metrics
- Proficiency in Python and deep learning frameworks (PyTorch, JAX, or similar) and experience with the Hugging Face ecosystem
- Experience building robust evaluation pipelines and a "scientific method" approach to understand which variables matter
- The ability to explain complex concepts to both engineers and executive stakeholders
- Proficiency in using coding agents to accelerate research and experimentation
- A Master's or PhD in CS/ML is a plus, but we value a track record of shipping production-grade ML systems above all else
Who we're looking for
- Unshakeable: High-agency executors who move fast but think deeply
- Curious: Quick learners who take time to listen, question, and explore
- Scientific but pragmatic: You care about the "why," but you prioritize what works in production
- Excellence: A passion for quality craftsmanship with a track record of top performance
- Resourcefulness: Proactive, independent, and able to drive projects forward from day one
What we offer
- Work on foundational infrastructure for the next generation of AI systems
- A small, highly capable engineering team where your work has immediate impact
- Direct exposure to cutting-edge enterprise AI deployments with global customers
- An office-centric culture that prioritizes output and responsibility over rigid schedules
- Competitive salary and equity stake
To apply
Ready to build retrieval infrastructure for the AI era? Send your CV and a brief cover letter about why you are a great fit for this role.
