Research Engineer (AI)

Location: San Francisco, CA, USA; on-site. Relocation assistance is available for those joining in San Francisco.

Salary: USD 180,000–250,000 per year plus equity.

Company

hirify.global is building an AI-human translation layer and multimodal foundation models focused on human goals, cognition, emotional wellness, and understanding people.

What you will do

• Orchestrate and optimize distributed training runs on long-horizon multimodal sequences, including DDP/FSDP configuration, mixed precision, checkpointing, recovery, and profiling.
• Build high-throughput data loading and processing pipelines for TB-scale multimodal data using sharding, prefetching, caching, and appropriate storage formats.
• Align and synchronize multi-stream time series, manage changing collection protocols, and maintain dataset versioning and lineage.
• Build reproducible feature-generation pipelines from raw recordings and ensure derived datasets are versioned and efficient to recompute.
• Develop clean, tested, documented research libraries, dataloaders, transforms, metrics, and evaluation harnesses.
• Make experiments traceable and turn research prototypes into repeatable pipelines while preserving research flexibility.

Requirements

• 3+ years of experience building ML systems or research infrastructure, including distributed training runs on 100+ GPUs.
• Expert-level Python and deep knowledge of PyTorch internals, including DDP, FSDP, mixed precision, gradient accumulation, and profiling tools.
• Track record of building or maintaining research libraries used by others, such as torch_geometric, torchaudio, torcheeg, torch_brain, neuralsets, or comparable internal tooling.
• Experience building data pipelines over large unstructured and multimodal datasets, with familiarity with Zarr, Parquet, Arrow, Lance, WebDataset, Vortex, and random-access versus sequential-throughput tradeoffs.
• Hands-on experience with experiment tracking and dataset versioning tools such as ClearML, Weights & Biases, MLflow, or similar.
• Research fluency sufficient to read a paper, reimplement a component, and assess whether a loss curve is broken.
• Strong debugging ability across corrupted shards, incorrect collate functions, and silently diverging multi-day training runs.
• Care for the well-being of people affected by the technology and ability to thrive in a high-bandwidth, collaborative environment.

Nice to have

• Custom kernel work in CUDA or Triton, or compiler-level optimization with torch.compile, TensorRT, or ONNX.
• Experience with time-series or multimodal data requiring synchronization across streams.
• Experience with low-latency or edge inference, quantization, or distillation.
• Comfort working in Rust or C++ when Python is not the right tool.
• Interest in neurotech, mental health, or human-AI interaction.

Culture & Benefits

• On-site work from a renovated office in San Francisco's Mission District with a full kitchen.
• Comprehensive health, dental, and vision insurance with premiums fully covered.
• Workspace setup stipend, relocation assistance, stocked kitchen, and dedicated meal budget.
• Bi-monthly team dinners and outings, plus twice-yearly off-sites and retreats.
• Conference travel sponsorship and regular professional development activities.
• At least 12 weeks of fully paid parental leave with a soft-landing transition, plus paid time off.