Skip to main content

Research Engineer, Latency & Throughput

Summary by AIEngineer.careers

Research Engineer, Latency & Throughput is a AI Research Engineer role (full-time). with Mistral.ai. in PARIS, FR. Imported listing (source: victra.careers). Apply on the employer's site (victra.careers).

Imported listing (source: victra.careers) · Apply on victra.careers

How our listings work

PARIS · FR
—
Full-time
MID
Apply Now

More AI jobs in France

The sections below reproduce the third-party job description for reference. AIEngineer.careers does not write or control this text.

Imported job description

Sourced from victra.careers

Mistral AI is seeking a Research Engineer for its Latency & Throughput team to advance LLM inference efficiency. You will research and prototype techniques such as speculative decoding, quantization, KV compression, and novel attention mechanisms, bridging academic research and production serving. The role involves close collaboration with inference engineers and research scientists to ship validated optimizations into a vLLM-based stack.

Responsibilities

  • Research and prototype inference techniques (speculative decoding, quantization, KV compression, watermarking, new architectures) by reviewing literature and forming testable hypotheses.
  • Design and run experiments to validate ideas, iterating quickly on small-scale setups and demonstrating latency/throughput gains without quality loss.
  • Collaborate with the Inference Foundation team to integrate validated techniques into the vLLM-based serving stack.
  • Partner with pretraining and RL teams when techniques require co-training, retraining, or auxiliary model training.
  • Build tooling and validation platforms to facilitate AI-driven inference research.

Requirements

  • Strong ML background with solid knowledge of LLM training techniques: architecture, tokenization, fine-tuning, and evaluation.
  • Hands-on experience with Python and PyTorch; JAX acceptable with strong Python proficiency.
  • Research experience (thesis, publications, dedicated project, or efficiency-oriented work in lab/industry).
  • Track record of conducting experiments: forming hypotheses, designing controlled comparisons, drawing clear conclusions.
  • Familiarity with state-of-the-art LLM inference techniques and open problems.
  • Self-starter, low-ego, collaborative, comfortable with ambiguity and negative results.

Nice to Have

  • Inference or serving experience (e.g., vLLM, production LLM deployments).
  • CUDA and GPU knowledge (hands-on not required).

Benefits

Health insuranceParental leaveretirement planrelocation supportwellness programmeal allowancetransportation allowance

Tech Stack

CUDAJAXLLM InferencePythonPyTorchQuantizationspeculative decodingvLLM
Mistral.ai

Mistral.ai

Paris · FR · 1019+ employees

Mistral AI is a French artificial-intelligence company founded in 2023 and headquartered in Paris. It develops open-weight and commercial large language models and related generative-AI products, with a mission to make frontier AI broadly accessible.

Weekly job alert

Get a weekly email when new AI engineering jobs match your filters. Free — confirm your address to start.

Related Jobs