ML Research Engineer, Performance Optimization

Lila Sciences; Cambridge, San Francisco (MA, CA); On-site; USD 148,000 - 240,000; Full time

Benefits offered by this job

Equity

Bonus potential

Comprehensive benefits

Office lunch program

Job summary

Lila Sciences in Cambridge, MA (US) seeks an ML Research Engineer focused on performance optimization for 1T+ parameter LLMs. You will maximize GPU utilization, profile kernels, and optimize cross-node throughput on large GPU clusters, aiming for scalable, maintainable improvements.

Responsibilities include RL pipeline optimization, data handling, and establishing a high-velocity, testable stack for researchers.

Qualifications

• BS/MS in Computer Science or related quantitative field and 4+ years as software or ML engineer, or PhD with 1+ year experience.
• Experience with LLM inference libraries such as vLLM and Megatron, including throughput optimization.
• Experience running LLM inference and training at scale across multi-node GPU clusters.
• Familiarity with reinforcement learning concepts.
• Proven track record of clean APIs, well-tested code, and reusable abstractions.

Responsibilities

• GPU compute performance profiling and kernel development.
• GPU cluster optimization, RDMA/IB/EFA usage, MoE framework tuning.
• Manage mid-training data across storage, streaming, tokenization, processing.
• Optimize RL pipeline and weight synchronization in async cycles.
• Ensure configurations are testable, accessible, and scalable for research tasks.

Skills

LLM optimization

GPU performance

CI/CD pipelines

Software engineering

Education

BS/MS in Computer Science or related field

Tools

vLLM

Megatron

CUDA

CuTe

CuTile

Triton

NCCL/NIXL

NVLink/IB/EFA

Ray

Job description

ML Research Engineer, Performance Optimization

Cambridge, MA USA; San Francisco, CA USA

Lila is building a platform where AI and automation co-evolve to accelerate scientific discovery Our AI Research team is looking to hire a ML Research Engineer to maximize iteration speed for mid- and post-training runs on 1T+ parameter LLMs using the latest NVIDIA chips. This involves not only maximizing GPU utilization on an individual training run but also making these optimizations maintainable, reusable, and adaptable so that the rest of the research team always knows how to confidently use the latest optimized stack.

What You'll Be Building

• GPU compute performance: profiling, kernel development, long-context handling, memory footprint optimization
• GPU cluster performance: collective/MoE framework optimization, comm/compute overlap, RDMA/IB/EFA usage, P/D disaggregation
• Big data handling: manage mid-training scale data across storage, streaming, tokenization, and processing
• RL pipeline: optimize the async RL cycle, including environment/sandbox interfacing, long rollout handling, large token metadata (such as router replay), and weight sync
• Setting the performance standard: make sure optimized configurations are tested, accessible, and adaptable to research tasks so that the whole team can work at maximum velocity
• Establishing the Pareto frontier: communicate with researchers about throughput/accuracy tradeoffs (e.g. lower-precision compute, higher async staleness) to establish optimal points

What You'll Need to Succeed

• BS/MS in Computer Science or a related quantitative field and 4+ years working as a software engineer or machine learning engineer, or PhD in a relevant research area and 1+ years work experience.
• Experience with LLM inference libraries such as vLLM and training libraries such as Megatron, including usage and throughput optimization.
• Experience running LLM inference and/or training at scale across multi-node GPU clusters.
• Familiarity with core concepts in reinforcement learning.
• Proven track record of software engineering excellence — clean APIs, well-tested code, thoughtful abstractions.
• Hands-on experience supporting and maintaining production libraries used by other engineers and researchers.
• Strong CI/CD skills — experience designing integration test suites, build pipelines, and automated quality gates for multi-package ecosystems.
• A passion for agentic systems and awareness of cutting-edge approaches in LLM tool use and autonomy.

Bonus Points For

• Experience working with kernel DSLs and frameworks like CUDA, CuTe, CuTile, Triton, etc.
• Experience with GPU networking frameworks like NCCL and NIXL, as well as fabrics like NVLink, IB, and EFA.
• Experience with distributed systems or compute frameworks (Ray, Flyte, Airflow, or similar).
• Familiarity with containerized compute backends (Docker, Modal) and sandbox orchestration.
• Experience with Ray actors, Ray Serve, or similar actor-model frameworks.

Compensation

We offer competitive base compensation with bonus potential and generous early-stage equity. Your final offer will reflect your background, expertise, and expected impact.

U.S. Benefits. Full-time U.S. employees receive a comprehensive benefits program including medical, dental, and vision coverage; employer-paid life and disability insurance; flexible time off with generous company wide holidays; paid parental leave; an educational assistance program; commuter benefits, including bike share memberships for office based employees; and a company subsidized lunch program.

International Benefits. Full-time employees outside the U.S. receive a comprehensive benefits program tailored to their region. USD salary ranges apply only to U.S.-based positions; international salaries are set to local market.

Expected Base Salary Range

$148,000 - $240,000 USD

About LILA

Lila Sciences is building Scientific Superintelligence to solve humankind's greatest challenges. We believe science is the most inspiring frontier for AI. Rather than hard-coding expert knowledge into tools, LILA builds systems that can learn for themselves.

LILA combines advanced AI models with proprietary AI Science Factory instruments into an operating system for science that executes the entire scientific method autonomously, accelerating discovery at unprecedented speed, scale, and impact across medicine, materials, and energy. Learn more at www.lila.ai.

Guided by our core values of truth, trust, curiosity, grit, and velocity, we move with startup speed while tackling problems of historic importance. If this sounds like an environment you'd love to work in, even if you don't meet every qualification listed above, we encourage you to apply.

Lila Sciences iscommitted to equal employment opportunityregardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender identity or Veteran status.

Information you provide during your application process will be handled in accordance with our Candidate Privacy Policy .