Senior AI Researcher (Post-Training)

Senior AI Researcher (Post-Training) Boston, Massachusetts (Hybrid / Remote) $180,000 - $250,000 + Equity + Healthcare + PTO Are you a Senior AI Researcher with a background in post-training language models looking to lead original research on autonomous, long-horizon agent systems within a high-growth cybersecurity startup? This is an opportunity to join a fast-growing AI company developing next-generation autonomous offensive security platforms. You will become the company’s first dedicated AI Researcher, working alongside engineering leadership and a specialized AI team to advance proprietary core models that reason, plan, and operate autonomously. In this role, you will drive original research on offensive security agents, focusing heavily on post-training pipelines, reinforcement learning from verifier signals, and co-evolutionary self-training architectures. This is a pure research role with a fast feedback loop, where you will design end-to-end experiments and translate research wins directly into production-ready capabilities. This role would suit an experienced AI Researcher who excels at building rigorous evaluation infrastructure from scratch and wants to pioneer how agentic systems behave under adversarial conditions. The Role: *Drive original research on agent reasoning, planning, and long-horizon execution *Advance post-training pipelines using supervised fine-tuning and reinforcement learning *Extend co-evolutionary self-training architectures and reward modeling systems *Design end-to-end experiments and build custom evaluation harnesses *Translate research breakthroughs into production handoffs for the engineering team The Person: *Proven track record of original machine learning research output or shipped systems *Hands-on post-training experience with large-scale language models (7B+ parameters) *Experience with reinforcement learning from verifier signals or preference optimization *Strong Python and PyTorch skills with multi-GPU distributed training experience *US Citizen Show more Show less