Senior AI Researcher

NextDeavor•New York, NY

6d•Hybrid

About The Position

Become a Key Player as a Senior AI Researcher. You will lead original research advancing core models that enable offensive-security capabilities, shaping experiments end-to-end and shipping results into production. You will collaborate closely with the VP of AI Engineering, the CEO, and a small AI engineering team to turn research outcomes into deployable capabilities.

Requirements

Demonstrated original ML research output (published papers, widely cited preprints, significant OSS releases, or shipped research that materially advanced a production system)
Hands-on post-training experience with large language models (7B+ parameters) and end-to-end ownership of data, training, and evaluation pipelines
Direct experience with at least one of: RL from verifier/reward signals, preference optimization (DPO/IPO/KTO), or supervised fine-tuning with synthetic data pipelines
Experience with agentic LLM systems: tool use, multi-step reasoning, planning, or long-horizon execution
Ability to design evaluations that measure real capability and avoid contamination or specification gaming
Strong Python and PyTorch skills, with experience in distributed multi-GPU training
Clear technical writing demonstrated by research memos, experiment writeups, or papers

Nice To Haves

Working knowledge of offensive security fundamentals (trainable on the job)
Prior work on code-generating or code-reasoning models
Experience with sparse, delayed, or expensive reward signals in RL
Research in robustness, adversarial ML, or red-teaming of language models
Familiarity with long-horizon agent benchmarks (e.g., SWE-bench, Cybench, WebArena)

Responsibilities

Drive original research on offensive-security agents: reasoning, planning, tool use, and long-horizon autonomous operation
Advance the post-training pipeline, including supervised fine-tuning, RL from verifier signals, LoRA adaptation, and adversarial evaluation
Extend co-evolutionary self-training architecture with curriculum design, self-play dynamics, and reward modeling for security outcomes
Design and execute experiments end-to-end, from hypothesis through writeup
Build internal evaluation harnesses where no public benchmark exists and measure capability rigorously
Translate research into production handoffs: model cards, deployment notes, and documented failure modes
Contribute to external research outputs: papers, talks, responsible disclosures, and technical writing
Collaborate with engineering teammates on research methodology and experimental design