Algorithm Engineer Intern

Algorithm Engineer InternCommitment Period: Jan 2027 onwards (minimum 6 months, full time)Job Description:We are looking for an Algorithm Engineer Intern to build search methods that find best compressed neural networks. Focus of this work includes neural architecture search (NAS) and on reinforcement learning (RL) for neural networks.Deliverable: reusable NAS tooling, and a comparison of RL against simple baselines for mixed precision.

Responsibilities

Enable common NAS patterns: search spaces and search strategies in reusable toolingExplore RL for mixed precision: per-layer bit-width selection, compared against simple baselinesBuild evaluation scripts with simple latency and size proxies, and report findingsTechniques You Will Work On:Common NAS techniques like search-space designRL: policy-gradient and actor-critic methods, reward designMixed-precision quantizationMinimum Requirements:Currently studying CS, Engineering, AI/ML or a related fieldCommitment for at least 6 months full timeStrong Python and hands-on PyTorchTrained and evaluated vision models end to endKnowledge of RL basics (MDPs, policy gradients)Knowledge of quantization basicsPreferred (Plus):Hands-on NAS or AutoMLExperience with Computer Vision modelsPublication or thesis in AutoML, RL or model compressionPublications at leading conferences (i.e. CVPR, ICML, NeurIPS etc.)