Senior ML Scientist – Edge AI & LLMs (Remote)

Attis; United States; Remote; USD 250,000 - 400,000; Full time

Benefits offered by this job

Equity

Remote work (USA)

Job summary

Attis is seeking a Machine Learning Engineer to build and optimize state-of-the-art LLMs for edge-ready deployments. You will drive post-training, alignment, and model compression across open-weight models, experimenting with quantization, distillation and advanced optimization techniques.

This fully remote role in the United States offers a salary between $250k and $400k plus equity. The ideal candidate has experience distributing training across GPU clusters, serving KV-cache-optimized

Qualifications

• Hands-on experience fine-tuning open-weight models with alignment techniques (GRPO, DPO, RLVR) or model compression methods (Quantization, Distillation, FP8/AWQ).
• Experience training across distributed GPU clusters (16+ H100/B200 nodes) using Slurm, FSDP, or DeepSpeed.
• Expertise in serving, parallel worker rollouts, and KV-cache optimization via vLLM, SGLang, or TensorRT-LLM.
• Exposure to mission-critical applications in Robotics, Tactical Edge AI, Autonomous Systems, or Defense preferred.

Responsibilities

• Develop and optimize end-to-end LLM post-training and alignment workflows.
• Experiment with model adaptation, compression, and evaluation to improve edge deployment.
• Collaborate with cross-functional teams to deploy scalable, secure AI solutions.

Skills

LLM

Reinforcement Learning

Post-Training

Model Compression

Distributed Training

Quantization

vLLM

PyTorch

Tools

Slurm

DeepSpeed

TensorRT-LLM

PyTorch

Job description

Attis is seeking a Machine Learning Engineer to build and optimize state-of-the-art LLMs for edge-ready deployments. You will drive post-training, alignment, and model compression across open-weight models, experimenting with quantization, distillation and advanced optimization techniques.

This fully remote role in the United States offers a salary between $250k and $400k plus equity. The ideal candidate has experience distributing training across GPU clusters, serving KV-cache-optimized