Are you passionate about building scalable, high-impact applications? At First Factory, we’re looking for a Computer Vision & Applied ML Engineer to join our team and work with a leading healthcare client on cutting-edge solutions. We would love to hear from you.

About First Factory

We are a software development company with over two decades of experience, boasting a dynamic team of 175+ professionals actively engaged in diverse projects across various industries. We invite you to join us on this journey as we thrive and embrace fresh challenges.

Key Responsibilities

• Own parts of an end-to-end pipeline from raw video through training, evaluation, and production inference, including dataset practices, metrics, and regression analysis.

• Train and iterate on computer vision models for objects, geometry, and human motion as they apply to a fixed, instrumented workflow (multi-angle video, occlusion, lighting variance).

• Curate and annotate video data, define evaluation protocols, and drive continuous improvement under deployment-like conditions—not leaderboard-only benchmarks.

• Design and maintain structured, time-aware representations of what the system observes (events, summaries, or sequences suitable for downstream reasoning and integration)—with attention to schema stability, normalization, and temporal alignment across cameras where relevant.

• Work with AWS SageMaker and Bedrock (and supporting services) for training, hosting, monitoring, and foundation-model workflows (inference, customization, and evaluation as we adopt them).

• Partner with hardware, backend, and product to integrate ML outputs into the broader platform safely and clearly.

• Contribute to model governance appropriate for healthcare: versioning, traceability, bias and error analysis, and documentation that supports compliance conversations.

Requirements

• 3+ years of hands-on computer vision and applied ML, with strong Python.

• PyTorch (or comparable framework) for training, fine-tuning, and debugging deep models, not only calling pretrained APIs.

• Practical video ML: frame sampling, temporal aggregation, multi-view considerations, and inference performance (latency, batching, basic optimization).

• Experience shipping or operating ML on AWS, including SageMaker (e.g., training jobs, endpoints, or batch transforms) and Bedrock for foundation-model use cases (inference and/or customization workflows you can explain concretely).

• Comfort with the full ML lifecycle: dataset versioning, experiment tracking, evaluation design, error analysis, and iteration with stakeholders.

• Hands-on experience building LLM applications with LangChain and LangGraph, designing chains, stateful graphs, tool/function calling, and multi-step reasoning flows around the fine-tuned model.

• Ability to collaborate in clear, written, and spoken English with a distributed team.

Nice to have

• Structured or JSON-centric ML interfaces: designing inputs/outputs for downstream reasoning or integration (without requiring public disclosure of our exact prompt or schema).

• Time-series or event-sequence thinking: aggregating noisy signals into stable behavioral summaries.

• Human action recognition, behavior analysis, or anomaly detection in video.

• Healthcare or regulated environments: PHI/HIPAA awareness, audit trails, change control.

• Annotation tooling (CVAT, Label Studio, Roboflow, etc.) and active learning or human-in-the-loop iteration.

• Relational data and services around ML (e.g., PostgreSQL) for metadata, jobs, or outcomes