CareerPlan

Working Student - LLM Quantization & Inference Performance

MunichRemoteFull-time2026-10-14 → 2026-11-16
Hybrid

Core

Evaluate LLM quantization methods to find the optimal tradeoff between model quality and inference runtime performance on custom silicon.

Role type

Working Student, LLM Quantization & Inference Researcher

Builds

Reusable evaluation tooling and quantization experiments for generative AI inference

Domain

Deep Tech / Generative AI / Hardware Acceleration

Deliverable

production ML models

Required skills

Python, PyTorch, LLM evaluation, experimental design, performance profiling

Preferred skills

LLM quantization, inference optimization, performance profiling

Technologies

PyTorch, Qwen, LLaMa

Responsibilities

Measure low-precision computation impact on coding and reasoning benchmarks; Implement and compare post-training quantization methods; Automate evaluations and track configurations; Characterize latency, throughput, and memory usage

Seniority

Student (Master's level)