Machine Learning Engineer (Detect & Track Distillation)

Harmattan AI
Paris, Zurich
Workplace: OnsiteFull timeFunction: Research & Scientific (R&D)Skills: ["Communication","Analytical thinking","Research orientation","Collaboration","Mentorship"]

Build and distill large foundation models into efficient, task-specific components for target detection, classification, and target re-identification. Optimize neural networks for constrained edge and embedded hardware using quantization, pruning, and LoRA, and develop end-to-end training and MLOps pipelines with strong reproducibility and logging. Drive benchmarking and evaluation against real-world latency and performance needs while collaborating with system and mission intelligence teams.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Harmattan AI
Harmattan AI
2 months ago

Machine Learning Engineer (Detect & Track Distillation)

✓ Verified Job

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

Source: Company careers pageValidated by: Fursa AI
Last checked: 16 hours agoStatus: Live

Job Summary

Build and distill large foundation models into efficient, task-specific components for target detection, classification, and target re-identification. Optimize neural networks for constrained edge and embedded hardware using quantization, pruning, and LoRA, and develop end-to-end training and MLOps pipelines with strong reproducibility and logging. Drive benchmarking and evaluation against real-world latency and performance needs while collaborating with system and mission intelligence teams.
Location: Paris, Zurich
Workplace: Onsite
Employment Type: Full time
Job Function: Research & Scientific (R&D)

Key Responsibilities

  • •Distill and fine-tune large foundation models into efficient components optimized for target detection and smaller tasks.
  • •Optimize neural networks for constrained embedded systems using quantization (PTQ vs. QAT), pruning, and LoRA.
  • •Build and maintain training, evaluation, and MLOps pipelines with reproducibility, robust logging, and version control.
  • •Curate data and develop task-specific datasets to improve model accuracy.
  • •Benchmark and evaluate distilled models for performance and latency aligned with real-world deployments, collaborating cross-functionally and mentoring junior engineers as needed.

Key Requirements

  • •Degree in a STEM field (e.g., Computer Science, Engineering, Mathematics).
  • •Proven experience running vision neural networks and developing target detection architectures or re-identification tasks.
  • •Hands-on expertise in knowledge distillation and model compression for deployment on constrained edge/embedded systems (e.g., Jetson, custom NPUs, wearables).
  • •Proficiency in MLOps with GPU compute and building training/evaluation infrastructure such as pipeline templates and loggers.
  • •Strong analytical, structured approach and ability to communicate complex benchmarking results to downstream users and senior stakeholders.
Education:
Skills:CommunicationAnalytical thinkingResearch orientationCollaborationMentorship
Tech Stack:Model distillationFine-tuningQuantizationPTQQATPruningLoRAMLOpsGPU computeVersion controlJetsonCustom NPUsWearablesComputer visionNeural networks

Company Brief

Harmattan AI
Develops AI-driven autonomous military systems (drones, electronic warfare, ISR and mission-management software) and scales conflict-ready production for allied armed forces and defense primes.
Industry: Defense Technology
Company Size: Medium (51 to 250 employees)
Growth: Scaleup
Valuation: Unicorn (USD 1B+)
Funding: Series B
Headquarters: Paris, France
Founded: 2024
WebsiteLinkedInGlassdoor