---
title: "Research Engineer, Robot Reinforcement Learning & Sim-to-Real"
slug: "research-engineer-robot-reinforcement-learning-sim"
id: "3a06548d-1976-4b70-ac23-f80510e6f6ab"
department: "Robotics & Embodied AI"
department_code: "ROB"
location: "San Francisco, CA"
workplace_type: "On-site"
employment_type: "Full-time"
compensation: "$235k – $330k (annual package)"
is_remote: false
is_featured: false
headcount: 1
canonical_url: "https://jobs.talent247.co/jobs/research-engineer-robot-reinforcement-learning-sim"
markdown_url: "https://jobs.talent247.co/jobs/research-engineer-robot-reinforcement-learning-sim.md"
apply_endpoint: "https://jobs.talent247.co/api/public/jobs/research-engineer-robot-reinforcement-learning-sim/apply"
schema_endpoint: "https://jobs.talent247.co/api/public/jobs/research-engineer-robot-reinforcement-learning-sim/form-schema"
webmcp_tool: "apply_for_role"
updated_at: "2026-11-15T17:34:39.837Z"
content_signal: "search=yes, ai-input=yes, ai-train=no"
---

# Research Engineer, Robot Reinforcement Learning & Sim-to-Real

## Position Telemetry

- **Reference:** `RESEARCH-ENGINEER-`
- **Practice / Sector:** Robotics & Embodied AI (`ROB`)
- **Location:** San Francisco, CA
- **Workplace Type:** On-site
- **Employment Type:** Full-time
- **Compensation:** $235k – $330k (annual package)
- **Canonical HTML URL:** https://jobs.talent247.co/jobs/research-engineer-robot-reinforcement-learning-sim

## Role Dossier & Specification

THE CLIENT & MANDATE
Our client is a frontier physical-AI research laboratory in San Francisco developing generalist robot foundation models across diverse real-world manipulation and mobile tasks. While human teleoperation provides broad behavioral priors, achieving industrial-grade speed, contact-rich precision, and autonomous recovery from mistakes requires scalable reinforcement learning across GPU-accelerated simulation and autonomous real-world practice.

They have engaged Talent247 to recruit a Research Engineer specializing in Robot Reinforcement Learning & Sim-to-Real onsite in San Francisco. The successful candidate will build massively parallel GPU physics gyms and real-world offline-to-online RL pipelines that push pretrained VLA policies far beyond demonstration performance.

KEY RESPONSIBILITIES
• Architect real-world and hybrid reinforcement learning algorithms (offline-to-online RL, Q-chunking, advantage-weighted flow matching, residual RL) fine-tuning large VLA policies directly on robot fleets.
• Build high-throughput GPU-accelerated physics simulation environments in MuJoCo MJX and NVIDIA Isaac Lab featuring automated real-to-sim asset reconstruction and contact-parameter system identification.
• Design autonomous reset mechanisms, VLM-based dense success/failure reward detectors, and safe exploration controllers enabling unattended multi-hour real-robot RL training runs.
• Engineer distributed actor-learner infrastructure streaming real-time multimodal rollouts from dozens of physical robot cells directly into centralized GPU training clusters.
• Investigate domain randomization, tactile/force-conditioned policies, and privileged-critic asymmetric architectures for contact-rich assembly and deformable-object manipulation.
• Collaborate closely with foundation model scientists to fold RL-refined high-skill trajectories back into generalist pre-training mixtures.

TECHNICAL ENVIRONMENT & OFFERING
Stack: Python, JAX, PyTorch, MuJoCo / MJX, Isaac Lab / Warp, C++, distributed RL actor-learner frameworks, multi-arm physical robot fleet & H100 GPU clusters. Workplace: Onsite in San Francisco, CA. Compensation: $235,000 – $330,000 base salary plus high-upside equity and comprehensive benefits.

## Core Requirements

ESSENTIAL QUALIFICATIONS
• MS or PhD in Robotics, Machine Learning, or Computer Science (or 4+ years of research engineering experience) with a strong focus on robot RL and control.
• Demonstrated success training RL policies that transfer to real physical robots (locomotion, dexterous manipulation, or contact-rich assembly) via sim-to-real or real-world RL.
• Deep proficiency in JAX or PyTorch alongside GPU-parallel physics simulators such as MuJoCo MJX, Isaac Lab/Gym, or Brax.
• Strong systems engineering ability to build low-overhead distributed rollout collection, replay buffers, and VLM reward-model pipelines.
• Solid foundations in off-policy RL, actor-critic stability, value overestimation mitigation, and continuous action distributions.

PREFERRED / BONUS EXPERIENCE
• Experience fine-tuning diffusion or flow-matching visuomotor policies with reinforcement learning (e.g., DPPO, IQL/Cal-QL, advantage-weighted regression, RLPD).
• Track record of publications at CoRL, RSS, ICRA, IROS, NeurIPS, or ICML.
• Hands-on familiarity with automated 3D scene/URDF reconstruction (3DGS/mesh-to-sim) for scalable task generation.

## Application Schema & Direct Apply (`apply_for_role`)

Candidates and authorized AI agents on behalf of a candidate can submit an application via the declarative WebMCP form tool `apply_for_role` on https://jobs.talent247.co/jobs/research-engineer-robot-reinforcement-learning-sim or via HTTP `POST https://jobs.talent247.co/api/public/jobs/research-engineer-robot-reinforcement-learning-sim/apply` (`application/json` or `multipart/form-data`). Non-browser HTTP agents must include header `X-Talent247-Intake-Token` obtained from `GET https://jobs.talent247.co/api/public/intake-token` (or automatically returned in `GET https://jobs.talent247.co/api/public/jobs/research-engineer-robot-reinforcement-learning-sim/form-schema`).

### Form Fields

- `full_name` (text, required): Full name
- `email` (email, required): Email
- `phone` (phone, optional): Phone
- `location` (text, optional): Location
- `linkedin_url` (url, optional): LinkedIn URL
- `cover_letter` (textarea, optional): Cover letter
- `consent` (checkbox, required): I consent to my data being processed for this application
- `cv_file` / `cv_base64` + `cv_filename`: Upload a CV in .pdf or .docx format (<= 10 MB) or provide an `https://` LinkedIn profile URL in `linkedin_url`.

### Alternative: Talent Pool Intake (`join_talent_pool`)

If this mandate is not an exact match, register once in the Talent247 Talent Pool at https://jobs.talent247.co/talent-pool (`https://jobs.talent247.co/talent-pool.md`) via WebMCP tool `join_talent_pool` or `POST https://jobs.talent247.co/api/public/talent-pool`.
