Skip to main content
TALENT247 // SPECIALIST AI EXECUTIVE SEARCH & TECHNICAL RECRUITMENT
RES · AI Research

Research Engineer, Post-Training & Reinforcement Learning

San Francisco, CA · Full-time · On-site

AI Specialism

AI Research

Workplace

On-site

Location

San Francisco, CA

Compensation

$265k – $370k

annual package

Role Dossier & Specification

REF: RESEARCH-ENGINEER-

THE CLIENT & MANDATE

Our client is one of the world's flagship frontier artificial intelligence research and deployment companies, headquartered in San Francisco, CA and serving hundreds of millions of global users and enterprise developers. Their Post-Training & Reinforcement Learning organization drives the reasoning, coding, mathematical problem-solving, and multimodal tool-use breakthroughs behind their flagship frontier models.

They have retained Talent247 to appoint a Research Engineer for Post-Training & Reinforcement Learning onsite in San Francisco. Embedded directly inside a tight-knit frontier post-training pod, the successful candidate will co-design large-scale online reinforcement learning algorithms, verifiable reasoning curricula, and asynchronous multi-thousand-GPU training infrastructure.

KEY RESPONSIBILITIES

  • Design, implement, and scale online reinforcement learning runs—spanning PPO, GRPO, and novel off-policy reasoning algorithms—across multi-hundred-billion-parameter dense and Mixture-of-Experts (MoE) architectures.
  • Architect high-throughput distributed rollout generation and verifier execution pipelines pairing low-latency inference engines with FSDP/Megatron training sharding at >50% MFU.
  • Construct verifiable and rubric-guided reward environments across competitive programming, multi-file software engineering, formal mathematics, scientific synthesis, and multi-turn agentic tool execution.
  • Diagnose and eliminate subtle post-training pathologies including entropy collapse, length hacking, sycophancy, verifier exploitation, and catastrophic forgetting of pre-trained world knowledge.
  • Run disciplined small-to-large scaling-law ablations that forecast test-time compute efficiency and downstream frontier benchmark jumps before committing cluster-scale compute.
  • Partner directly with pre-training, inference kernel, and safety alignment teams through flagship model release cycles.

TECHNICAL ENVIRONMENT & OFFERING

Stack: Python, PyTorch, Triton/CUDA, custom distributed RL actor-learner frameworks, Megatron/FSDP, Kubernetes, multi-thousand H100/B200 GPU superclusters. Workplace: Onsite in San Francisco, CA. Compensation: $265,000 – $370,000 base salary plus top-decile frontier lab equity, comprehensive medical coverage, daily meals, and full relocation/visa support.

Core Requirements

ESSENTIAL QUALIFICATIONS

  • 4+ years of research engineering or ML systems experience (or PhD in Machine Learning/CS) with direct ownership of large-scale LLM post-training or reinforcement learning pipelines.
  • Deep production mastery of PyTorch internals, distributed parallelism (TP, PP, EP, CP, FSDP), and numerical stability debugging in BF16/FP8 regimes.
  • Hands-on track record implementing RLHF, RLAIF, PPO, GRPO, DPO, or verifiable-reward RL curricula on transformer architectures.
  • Strong systems programming ability to profile GPU kernel bottlenecks, NCCL collective stalls, and asynchronous rollout-to-learner network throughput.
  • High experimental velocity and scientific rigor when isolating subtle algorithmic gains in noisy RL training curves.

PREFERRED / BONUS EXPERIENCE

  • Core contributions to shipped frontier reasoning models, competitive-coding/math RL breakthroughs, or major open-source RLHF frameworks.
  • Experience co-designing continuous-batching inference engines (vLLM/SGLang/TensorRT-LLM) for on-policy RL rollout generation.
  • Publications at NeurIPS, ICML, ICLR, or ACL, or strong competitive programming / Olympiad background.

Position Telemetry

Location San Francisco, CA
Department AI Research
Employment type Full-time
Workplace type On-site
Headcount 2 openings
Salary range
$265k – $370k annual package