Applied Research Engineer / Research Intern

Please note this is a remote role (Preferably India) with heavy equity for the right person.

About us

We are building always-on AI coworkers for the enterprise: multi-agent systems that hold governed memory, act in real time across a company's tools, and solve problems rather than just answer questions. We're an early-stage, well-capitalized team, which means you'll work directly with the founder, ship research into production quickly, and see your work reach real users in weeks rather than years.

The role

You'll sit at the boundary between research and production. You'll read the latest agent and LLM literature, reproduce what matters, prototype it against our architecture, and then harden what works into shipped systems. This is not a pure-publication role and it's not a pure-engineering role. You should be equally comfortable reimplementing a paper from scratch and debugging a latency regression in production.

What you'll work on

• Multi-agent orchestration: designing, implementing, and evaluating how multiple agents plan, delegate, and recover from failure on long-running enterprise tasks.
• Governed agent memory: retrieval, consolidation, and forgetting; provenance and access control so an agent only ever uses information it's permitted to see.
• Real-time inference: reducing end-to-end latency across the edge-cloud boundary through caching, speculative execution, streaming, and model routing.
• Evaluation: building the benchmarks and regression harnesses that tell us whether a change actually improved agent reliability, since this is where most agent work falls down.
• Tool use and integration: making agents reliably operate real enterprise APIs, with retries, verification, and safe failure modes.
• Open-source contribution: publishing packages and plugins that build a developer community around our work.

Minimum qualifications

You should meet one of the following:

• MS in Computer Science, Machine Learning, or a related field, plus 2+ years of research experience (industry research, a research lab, or published work).
• PhD in a related field, plus 1+ year of research experience.
• Currently 3+ years into a PhD program (for the internship track), with a track record of independent research.

You should also have:

• Strong Python. This is our primary language. You should write clean, tested, readable code, not just research scripts.
• Hands-on experience with modern ML frameworks, such as PyTorch, JAX, Hugging Face Transformers, and the current agent and LLM tooling ecosystem.
• Demonstrated research output: publications, preprints, an open-source project, a thesis, or a reproducible body of work you can walk us through.
• The ability to go from paper to prototype independently. We'll give you a problem and context, not a spec.

Nice to have

• Experience with LLM agent frameworks, retrieval systems, or vector databases.
• Work on inference optimization: quantization, KV-cache management, batching, or serving infrastructure.
• Distributed systems or edge deployment experience.
• Familiarity with enterprise constraints such as SOC 2, data residency, RBAC, and audit logging.
• Open-source maintainership or a strong public GitHub presence.
• Prior startup experience, or any evidence you thrive without much structure.

What we offer

• Direct work with the founder and real influence over the technical direction.
• A fast path from idea to production, with no research-to-engineering handoff wall.
• Meaningful equity (full-time), competitive compensation, and support for publishing your work.
• Interns get a defined project, a named mentor, and a serious shot at a full-time offer.