Summary

Machine Learning Scientist at Cellular Intelligence. Working on multi-modal foundation models for life sciences.

Education

• MSc Computer Science (Machine Learning)

  2024-09
  Imperial College London

  Courses: Deep Learning, Reinforcement Learning, Graph-Based Learning, Software Engineering for ML Systems, Natural Language Processing
• BEng Engineering (Robotics)

  2023-06
  Imperial College London

Work Experience

• Machine Learning Scientist

  2024-09 -
  Cellular Intelligence

  6th engineer at a Harvard-MIT spin-off building multi-modal foundation models for life sciences ($60M+ raised, led by Khosla Ventures).

  • Co-authored publications at ICML, ICLR, and Nature Machine Intelligence on pre-training strategies and measurement noise scaling laws for foundation models.
  • Designed and implemented pre-training and post-training pipelines for billion-parameter transformers on terabytes of data across 32 AMD MI325x GPUs using PyTorch FSDP and SLURM.
  • Deployed ML services to production on AWS: containerized React + FastAPI application with Docker, built MCP servers for model integration, integrated RAG pipelines for document retrieval, and served via Nginx.
  • Developed a reasoning LLM via rejection sampling for data curation, SFT for cold-start, and GRPO with vLLM rollouts using Hugging Face — improving automated biological annotation quality.

Research Experience

• Visiting Researcher

  2024-04 - 2024-09
  Harvard University (PI: Prof. Stephanie Gil)

  Designed efficient transformer attention mechanisms for spatio-temporal demand forecasting, applied to shuttle bus scheduling on the Harvard campus.
• Undergraduate Researcher

  2023-06 - 2023-08
  Caltech (PI: Prof. Aaron Ames)

  Developed CVAEs and diffusion models to learn residual dynamics for drones, enabling tighter safety guarantees with discrete-time control barrier functions; resulted in an ICRA 2024 publication.
• Undergraduate Researcher

  2022-06 - 2022-08
  Caltech (PI: Prof. Aaron Ames)

  Accelerated model predictive control solvers by implementing custom matrix operations, reducing per-step computation time for real-time robotic control.

Skills

Publications

• Large-Scale Benchmarking of Gene and Expression Encoding Strategies for Single-Cell Foundation Models

  2026
  ICLR 2026 Workshop on Generative AI for Genomics

  First author. Presented at ICLR 2026 Workshop on Generative AI for Genomics.

  View Publication
• Scaling Laws for Noise for Cellular Representation Learning
  Nature Machine Intelligence (Under Review)

  Under review at Nature Machine Intelligence.

  View Publication
• Scaling Up Scaling Laws for Noise

  2025
  ICML 2025 Workshop on Multimodal Foundational Models for Life Sciences

  Presented at ICML 2025 Workshop on Multi-modal Foundational Models for Life Sciences.

  View Publication
• Generative Modelling of Residuals for Real-Time Risk-Sensitive Safety with Discrete-Time Control Barrier Functions

  2024
  IEEE International Conference on Robotics and Automation (ICRA) 2024

  Published at ICRA 2024.

  View Publication

Teaching

• Polish AI Olympiad
• Open Avenues
• Imperial Driverless