Intern Researcher - Agentic ML Compiler

Description du poste

yasp is pioneering the future of software development with a compiler that leverages agentic AI for advanced optimization and code generation. We are looking for an Intern Researcher - Agentic ML Compiler to join our team and help drive the core innovation that will define the next generation of our technology.

We don’t draw boundaries between research and engineering. We are creative thinkers and relentless prototypers who live at the bleeding edge of AI research and engineering. Together, we explore, invent, and ship novel technology that pushes beyond the state of the art for accelerating deep learning models.

Tâches

• Develop novel agentic AI and LLM-based methods to optimize kernels for machine learning models.
• Evaluate the performance of AI driven ML compilation techniques via rigorous benchmarking.
• Stay up to date with the publication landscape, open-source projects, and industry trends.
• Contribute to writing research papers for publication in top AI conferences.
• Work with a talented team towards breakthroughs in AI for ML compilation.

Exigences

• Master’s or PhD student in Computer Science/Maths/Physics with a focus on AI/ML, or equivalent practical experience.
• Exceptional prior work in AI/ML, ideally involving LLMs or agents to solve complex tasks.
• Strong engineering and coding abilities (PyTorch, agentic AI frameworks, LLM post-training frameworks, software development best practices).
• Passion for good scientific practice, blending clean formal thinking and good communication with exacting engineering.

Préféré

• Experience post-training LLMs using SFT and RL, especially as part of an AI agent.
• Knowledge of ML compiler principles including intermediate representations, operator fusion, and custom CUDA or Triton kernel generation.
• Previous publications in leading AI conferences (e.g., NeurIPS, ICLR, ICML).

Ce que nous offrons

• Competitive salary.
• Comprehensive health benefits (including dental).
• Opportunities for professional development and growth.
• Flexible work hours.
• Dynamic and collaborative work environment.
• Cutting-edge software and hardware platforms.

Processus de candidature

Ready to apply? Submit your application through this posting and our team will review it personally. From there, you can expect a short intro call, a technical evaluation built around real problems, and conversations with the people you'd work with directly.

À propos de yasp

yasp is building toward a future where every AI workload runs on the optimal hardware automatically. Today, getting a model from research into fast, reliable production demands scarce specialists in compilers, kernels, and runtimes. yasp is an agentic AI optimization platform that hands that work to agents, so the labor that bottlenecks AI deployment today becomes invisible infrastructure tomorrow.

Our agents take on the iteration, exploration, and tuning that consume months of engineering time, working at a depth and persistence that is difficult to sustain by hand. We start where the pain is sharpest: helping ML and performance teams ship faster to production and extract more from the hardware they already run on, across Nvidia, AMD, and AWS Trainium. From there, the platform expands outward, workload by workload, hardware target by hardware target.

Combining deep compiler and systems expertise with agents that operate at machine speed, and working in close collaboration with hardware vendors, yasp is heading toward a world where deploying optimized AI stops being a technical problem at all. We serve real-time and edge AI companies (including autonomous vehicles, robotics, drones, aerospace, computer vision, and medical devices), hardware vendors, and enterprises running AI in production.