Enterprise AI Research Scientist — Long-Horizon Agents

Cerebras; San Jose (CA); On-site; USD 140,000 - 210,000; Full time

Job summary

Tessera Labs is seeking a Research Scientist to set and pursue a research agenda for reliable long-horizon agents operating inside real enterprises. You will invent methods, work with Research Engineers, and aim to publish discoveries alongside product impact.

You’ll explore verification, memory architectures, and evaluation methodologies while ensuring work ships with the product team in a fast-moving enterprise setting.

Qualifications

• Have an MS or PhD in CS, ML, statistics, math, physics, or a related field.
• Have a track record of original research in ML.
• Deep expertise in post-training/RL for LLMs, agent memory and long-context reasoning, knowledge representation, or evaluation methodologies.
• Hands-on: you write code, run experiments, and read logs.
• Hold rigorous judgment and urgency; pursue true results with clear ownership.

Responsibilities

• Set and pursue a research agenda on reliable long-horizon agentic behavior in real enterprise environments.
• Invent and validate methods for post-training agents: reward design, RL for long-horizon planning, curriculum and data strategy.
• Define memory architecture for multi-day runs: persistence, retrieval, revision.
• Develop evaluation methodology预测 customer-observed correctness and compare proxies.
• Study multi-agent failure modes and design against it.
• Address verification problems to ensure changes preserve behavior.
• Turn research into ship-ready artifacts with engineering and product.
• Publish papers, technical reports, and open-source artifacts, and represent Tessera's research externally.
• Mentor engineers and raise the team's research bar.

Skills

Research in ML

Post-training RL

Agent memory & long-context

Evaluation methodology

Knowledge representation

Education

MS or PhD in CS/ML/Statistics/Math/Physics

Job description

Tessera Labs is seeking a Research Scientist to set and pursue a research agenda for reliable long-horizon agents operating inside real enterprises. You will invent methods, work with Research Engineers, and aim to publish discoveries alongside product impact.

You’ll explore verification, memory architectures, and evaluation methodologies while ensuring work ships with the product team in a fast-moving enterprise setting.