Hatch logo for LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness)  @ job on DevsUnite

LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness) @

Hatch

Actively HiringFull-time
Apply
Location

Australia

Experience

3+ years

Role

LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness) @

Salary

Competitive

Share this job

Know someone who'd be a great fit?

Job Description

Role Overview

We're looking for experienced machine learning researchers with hands-on experience training and improving deep learning models end-to-end, across vision and language. You'll work on well-scoped empirical open-ended ML research problems.

Responsibilities

  • Train image classifiers and generative image models from scratch, and fine-tune open-weight language models.
  • Get the most out of limited data, compute, and model-size budgets.
  • Make models robust to adversarial inputs and adversarial conversations.
  • Compress models to meet hard size and latency constraints without sacrificing accuracy.
  • Diagnose and resolve training issues.

Requirements

  • Adversarial Robustness: PGD-based training, TRADES, L∞ attacks, AutoAttack, managing robustness-accuracy trade-off.
  • Efficient Computer Vision: End-to-end training, fine-grained recognition, quantization, pruning, knowledge distillation, edge deployment.
  • Generative Image Modeling: Diffusion models, GANs, VAEs, FID metrics, training-efficiency tricks.
  • LLM Post-Training & Behavioral Robustness: DPO, RLHF, RLAIF, multi-turn conversation robustness, alignment-style fine-tuning.
  • Multilingual Pre-training: Low-resource languages, tokenizer design, cross-lingual transfer.
  • General Qualifications: 3+ years ML research; strong experience with PyTorch, JAX, TensorFlow; top-100 university or FAANG equivalent.

Required Skills

PyTorchJAXTensorFlowDiffusion Models

Ready to apply for LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness) @?

See how you stack up at Hatch before you apply, or go straight to the application.

Apply Anyway