MLFF Distillation & GCMC Integration - Internship

Cusp.AI
London, Cambridge
Workplace: HybridInternshipFunction: Healthcare (Clinical, Medical, Wellness)Education: phdSkills: ["Dataset curation","Documentation","Benchmarking","Collaboration","Systems thinking"]

Develop fast, accurate machine-learning force fields (MLFFs) for high-throughput Monte Carlo and integrate them into the company’s in-house simulation framework, kUPS. Distill state-of-the-art equivariant models into lightweight student potentials, curate and document training/validation datasets, and run head-to-head validation versus classical force-field baselines. Profile and optimize the MC inner loop for throughput and contribute to research advancing MLFF-driven GCMC for MOF screening.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Cusp.AI
Cusp.AI
1 month ago

MLFF Distillation & GCMC Integration - Internship

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 3 hours agoStatus: Live

Job Summary

Develop fast, accurate machine-learning force fields (MLFFs) for high-throughput Monte Carlo and integrate them into the company’s in-house simulation framework, kUPS. Distill state-of-the-art equivariant models into lightweight student potentials, curate and document training/validation datasets, and run head-to-head validation versus classical force-field baselines. Profile and optimize the MC inner loop for throughput and contribute to research advancing MLFF-driven GCMC for MOF screening.
Location: London, Cambridge
Workplace: Hybrid
Employment Type: Internship · 3 months
Job Function: Healthcare (Clinical, Medical, Wellness)
Seniority: Intern level

Key Responsibilities

  • •Distill MLFFs into fast student potentials optimized for Monte Carlo simulations.
  • •Curate, version, and document training/validation datasets, including distillation protocol and any active-learning loops.
  • •Run head-to-head validation campaigns comparing distilled MLFFs against classical force-field baselines across curated guest molecules.
  • •Profile and optimize the simulation pipeline for throughput, focusing on the MC inner loop where inference cost dominates.
  • •Collaborate with computational chemists on reference data generation and contribute to publication work on MLFF-driven GCMC for MOF screening.

Pay and Benefits

Perks:EquityPaid LeaveParental LeaveLearning Budget

Key Requirements

  • •Currently enrolled in (or recently completed) a PhD or Master's in a relevant quantitative field such as Physics, Chemistry, Chemical Engineering, Computational Science, or Machine Learning.
  • •Experience in adsorption modelling at the atomic scale.
  • •Hands-on experience with molecular simulation methods, including GCMC and/or MD.
  • •Comfortable working on Linux and managing simulation campaigns at scale.
  • •Genuine interest in applying ML to chemistry and materials science.
Experience:Materials scienceComputational chemistryMachine learningGas adsorptionMOFs
Education:PhD / Doctorate in Physics, Chemistry, Chemical Engineering, Computational Science, Machine Learning (or similar)
Skills:Dataset curationDocumentationBenchmarkingCollaborationSystems thinking
Tech Stack:MLFFMonte CarloGCMCMDLinuxEquivariant modelsKnowledge distillationActive learningDFTMOFsClassical force fieldsKUPS

Company Brief

Cusp.AI
CuspAI builds AI foundation models and simulation tools to accelerate discovery and design of novel materials for applications like batteries, semiconductors, water treatment, and carbon capture, combining generative models with physics-based simulation.
Industry: Materials Science
Company Size: Small (11 to 50 employees)
Growth: Growth Stage Startup
Valuation: USD 500M to 1B
Funding: Series A
Headquarters: Cambridge, United Kingdom
Founded: 2024
WebsiteLinkedInGlassdoor