Search Jobs
Apply Now

The Opportunity

EQL Tech is partnering with PhormAI to hire its first engineer. PhormAI is building Darwin, a Physical AI platform for Health & Wellness that turns multimodal sensor data – wearables, video, IMU, ECG, EMR – into clinical-grade insight through agentic AI. It’s not a prototype: Darwin is live with paying customers in two distinct verticals today – equine cardiac monitoring (animal health) and tele-ICU intelligence across multiple hospitals (human health) – with the architecture built to extend into further verticals by swapping domain ontologies rather than rebuilding the stack.

You’d own significant portions of the Darwin platform end-to-end — the data and model pipelines, cloud infrastructure, the agentic AI runtime, and the application layer built on top. You work directly with all three founders on architecture and technical strategy.

The Role

You own the technical foundation of the product: data and model pipelines, cloud infrastructure, developer tooling, and the application layer built on top of it.

  • Data pipelines – build the systems that ingest, process, and transform data for the models.
  • Cloud infrastructure – stand up and operate compute, storage, networking, and deployment from scratch.
  • Orchestration – connect ingestion, training, inference, evaluation, and serving into one coherent pipeline.
  • Agentic runtime — build multi-agent orchestration (LangGraph or equivalent), an LLM gateway with cost/latency-aware model routing and semantic caching, RAG pipelines, and continuous evaluation for hallucination detection.
  • Model operations — operate self-hosted Physical AI models on AWS SageMaker, EKS, or ECS with GPU/CPU auto-scaling and Spot instance cost discipline.
  • Compliance infrastructure — establish multi-environment isolation using AWS Organizations, multi-VPC architecture, and IAM policies that hold up under HIPAA scrutiny.
  • Observability — implement real-time monitoring (CloudWatch, Datadog, Prometheus/Grafana) for API latency, throughput, and prompt/response telemetry.
  • CI/CD — build automated pipelines for zero-downtime deployment (blue/green or canary) of models, API wrappers, and containerised serving applications.
  • Application layer – contribute to the APIs, services, and user-facing surfaces built on top of the platform.
  • Technical strategy – partner with Data Science and Product on architecture; own decisions that show up in customer SOWs.

Requirements

What You’ll Need

5+ years of software engineering, with at least 1-2 years shipping GenAI or ML systems in production. Not every line below is a gate; what matters is first-principles understanding of the stack and a history of building end-to-end systems that made it to production and stayed there. If you’ve done the following in production, not just in a course or a side project, you’re in range:

  • Strong Python and SQL.
  • Built and operated a production data pipeline (Airflow, Dagster, Prefect, or similar).
  • Worked hands-on with a cloud platform (AWS, GCP, or Azure) – provisioning, IAM, networking.
  • Shipped with Docker, and either Kubernetes or a managed container platform.
  • Infrastructure-as-code in production (Terraform, Pulumi, or CDK) – you provision, you don’t just click around a console.
  • Designed and operated a backend API (REST, GraphQL, or gRPC) including auth.
  • Relational database design and query optimization at scale.
  • Real-time streaming (Kafka, Pub/Sub, Kinesis, Flink).
  • Used an AI coding tool (Claude Code, Cursor, Copilot, or similar) as part of your actual workflow.
  • A track record of debugging across a stack you didn’t fully build yourself.
  • FastAPI or equivalent async web framework
  • GenAI systems in production: multi-agent orchestration (LangGraph, LangChain), RAG architectures, prompt engineering, continuous evaluation (LangSmith or equivalent)
  • Vector stores: pgvector, OpenSearch, or Pinecone
  • Redis or equivalent for caching, deduplication, and low-latency read paths
  • Observability: Prometheus, Grafana, Datadog, CloudWatch — metrics, logs, and traces
  • CI/CD with zero-downtime deployment patterns (blue/green, canary) — currently implied by Docker/K8s bullet but not stated
  • API gateway configuration for high-throughput inference workloads — more specific than current “designed and operated a backend API”

Would help, not required:

  • Health, clinical, or regulated-data experience – genuinely not a gate; most of this team came from outside health.
  • Full-stack product experience – you’ve shipped a user-facing web app end to end.
  • Open-source contributions.
  • HIPAA, SOC 2, HITRUST, or GDPR experience
  • Self-hosted ML model deployment on SageMaker, EKS, or ECS with GPU inference and model artifact versioning
  • Biomedical signal processing (BioSPPy, Neurokit2, MNE-Python)
  • Computer vision pipelines (OpenCV, MediaPipe, YOLO, FFmpeg)
  • MLOps tooling: MLflow, Weights & Biases, LangSmith
  • Multi-tenant SaaS architecture at scale

Benefits

Compensation & Equity

  • Base salary $150,000 – $200,000+ depending on experience, paired with a strong, early equity grant. You’re getting in at the earliest stage of the engineering function — before the team scales and before the architecture is locked. Exact base and equity percentage confirmed directly with the founders during the process.

Why This One:

    • You’re not the fifth infra hire cleaning up someone else’s decisions – you’re making them.
    • The product already has paying customers in two verticals; you’re not evangelizing a concept, you’re scaling one that’s proven.
    • Small, high-trust team. No legacy debt to inherit.

Similar jobs