
Senior/Lead Research Engineer, Foundation Models
Posted 1 day ago

Posted 1 day ago
This is a fully remote position, open to applicants in United States, +1 more country.
• Develop and maintain the systems that support Thomson's training, evaluation, and experimentation processes, ensuring their reliability as model usage increases.
• Assess and optimize for cost efficiency, latency, and GPU utilization without compromising on quality, while documenting trade-offs for the broader team.
• Diagnose complex numerical, distributed, and performance issues across training frameworks, inference engines, and evaluation code.
• Collaborate directly with the training and evaluation teams to facilitate the training, assessment, and deployment of new model releases.
• Contribute to Thomson's production infrastructure, including the path to ISO 27001 compliance.
• Establish benchmarks for code quality, testing, and reproducibility.
• At the Lead level: take ownership of the architecture in your domain, strategize delivery and staffing in collaboration with research leads, and provide mentorship to engineers.
• Create systems that train, evaluate, and deploy Thomson's foundational models for professional applications.
• MSc or PhD in computer science, machine learning, or a related discipline, or equivalent experience in the industry.
• Strong foundation in software engineering practices, including large-scale distributed systems and proficiency with extensive open-source codebases.
• Proven experience in maintaining the reliability of research or production systems under real workloads, alongside debugging distributed systems and incidents.
• Experience collaborating closely with researchers and cross-functional stakeholders to transform research concepts and requirements into scalable systems.
• Expertise in at least one specialized area: Large-scale RL training involving multi-node GPU training with PyTorch and Megatron-Core, FSDP/DTensor, or DeepSpeed; parallelism strategies (tensor, pipeline, context, expert) and their implications for MoE and long-context models; LLM inference engines (vLLM, SGLang); GPU profiling (memory, NCCL, CUDA graphs). Evaluation engineering: designing, scaling, and optimizing evaluation suites for LLMs, along with demonstrable impacts of your contributions.
• For Lead level: Over 3 years of professional experience in your field, with a track record of leading complex technical projects from inception to completion and mentoring engineers.
• Experience building or enhancing RL post-training frameworks (NeMo-RL, verl, OpenRLHF, TRL), specifically with asynchronous or non-co-located designs.
• Familiarity with evaluation frameworks like Inspect, Harbor, or lm-evaluation-harness.
• Experience with distributed runtimes and orchestration tools (Ray, Kubernetes, Slurm) on platforms such as AWS, Azure, or Google Cloud.
• Significant contributions to vLLM, Megatron, PyTorch, or similar open-source initiatives.
• Familiarity with AI-assisted coding and the utilization of coding agents at scale.
• Flexible work arrangements, including the option to work from anywhere for up to 8 weeks each year.
• Flexible vacation policy.
• Two designated company-wide Mental Health Days off.
• Access to the Headspace app.
• Retirement savings plans.
• Tuition reimbursement opportunities.
• Employee incentive programs.
• Resources focused on mental, physical, and financial wellbeing.
• Two paid volunteer days off each year.
• Comprehensive health, dental, vision, disability, and life insurance programs.
• Competitive 401k plan with company matching contributions.
• Paid holidays.
• Parental leave policies.
• Sabbatical leave options.
• Optional hospital, accident, and sickness insurance fully paid by the employee.
• Optional life and AD&D insurance fully paid by the employee.
• Flexible Spending and Health Savings Accounts.
• Fitness reimbursement program.
• Access to Employee Assistance Program.
• Group Legal Identity Theft Protection benefit fully paid by the employee.
• Access to a 529 Plan.
• Commuter benefits.
• Adoption & Surrogacy Assistance.
• Access to Employee Stock Purchase Plan.
• Annual bonus based on a combination of enterprise and individual performance.
Dragos, Inc.
Dragos, Inc.
Ramp
Get handpicked remote jobs straight to your inbox weekly.