Machine Learning Engineer — Multimodal & Vision
AI Engineering · Machine Learning & AI · United States · New York, NY / Remote (USA) · Remote · Full-time · Mid-Senior (3+ yrs) · $140,000 – $185,000 + Equity
Deploy assisted pre-labeling vision models and zero-shot architectures that dramatically accelerate throughput for autonomous driving and robotics datasets.
Responsibilities
- Fine-tune and deploy assisted pre-labeling vision models (SAM, YOLO, Mask R-CNN, 3D point cloud detectors)
- Optimize model inference latency on GPU clusters using NVIDIA TensorRT, ONNX, and Triton
- Evaluate model prediction confidence against human consensus to tighten the active learning loop
- Integrate model-assisted annotation APIs directly into customer MLOps infrastructure
Requirements
- 3+ years building and deploying deep learning models in production (PyTorch, Docker, CUDA)
- Proven expertise in Computer Vision, 3D LiDAR/sensor fusion, or multimodal vision-language architectures
- Strong software engineering fundamentals in Python and cloud systems (AWS/GCP)
- Experience optimizing deep learning inference for real-time interactive applications
Perks
- Dedicated cloud GPU compute budget for experimental research
- Full healthcare coverage including mental wellness stipends
- Flexible working hours across any global timezone
- Generous hardware allowance (M3 Max MacBook Pro or high-end Linux workstation)