Senior Software Engineer, Edge & Cloud Platform

Who We Are

illumicell AI is a medical-technology startup building AI-powered, label-free cell analysis to bring fast, objective diagnostics to the point of care. Based at the Pagliuca Harvard Life Lab, Harvard's startup incubator in Allston, we work at the intersection of optics, machine learning, and clinical medicine, surrounded by the research community our technology grew out of.

Starting with male fertility, we're developing a new standard for how living cells are measured, understood, and acted upon. We have a functional prototype device, a digital holographic scanner paired with cloud-based AI analysis, already producing live-cell measurements, and we are rapidly accelerating toward FDA submission. We're backed by Accel and Tim Draper, are Techstars alumni, and are building the team that will carry the product from working prototype to cleared clinical device.

About the Role

illumicell AI is building a holographic imaging system for automated semen analysis. The product begins with a Linux-based microscope scanner that captures live-cell video and ends with a clinically deliverable result: sperm concentration, motility, and eventually morphology, returned to the fertility lab. Between those two points is the platform this role is responsible for.

We are looking for a Senior Software Engineer to build and maintain the software that connects the scanner to cloud-based image reconstruction and machine learning analysis, and returns a reliable result to the clinic. That means device control and image acquisition, patient-safe data handling, resilient upload, cloud job orchestration, GPU analysis integration, result delivery, and the observability and test infrastructure to keep the whole path working.

This is a hands-on product role, not a DevOps or internal-tooling position. You will write production software across Linux-based edge devices and cloud services, define the interfaces between them, and carry difficult work through integration and verification. You do not need equal expertise in every layer, but you should be able to learn adjacent technologies and contribute wherever the work is most pressing.

You will work alongside an existing senior engineer who leads GCP, infrastructure, IAM, device-identity, deployment, and ML-infrastructure work. This role adds a second senior product-systems path and initially emphasizes edge and cloud application logic, interfaces, end-to-end integration, and shared review. It is not a replacement role. Level may flex to Staff for a candidate who has demonstrated technical direction across device, cloud, and infrastructure teams.

What You Will Build

The scanner captures ordered monochrome frames, removes patient identity before anything leaves the device, and uploads the encoded acquisition. The cloud verifies and admits the package, runs one GPU analysis job, and returns a de-identified result to the originating device, where identity is restored and the report is delivered to the clinic. Near-term work includes:

• A restart-safe edge controller and explicit scan state machine
• Camera acquisition running as a testable service rather than an interactive prototype
• Lossless PNG or FFV1 packaging, checksums, manifests, and resumable transfer to GCS
• An on-device patient-identity boundary and tokenized device-to-server contract
• Device-local REST and event-stream APIs for the operator interface and test automation
• An authenticated cloud control and admission API
• Event-driven cloud processing using GCS, Eventarc, Firestore, and Cloud Run Jobs
• Idempotent job launch, immutable result publication, retries, reconciliation, and cleanup
• Model and configuration provenance across reconstruction and ML analysis steps
• CI, structured observability, security controls, dependency records, and regulated traceability

Responsibilities

• Take responsibility for major edge and cloud product capabilities from problem definition through deployed and tested behavior.
• Build services and libraries for Linux edge devices with explicit state, resilient local storage, and hardware-aware failure handling.
• Design and implement versioned REST, JSON, event, and storage contracts between the scanner, operator interface, cloud control plane, and analysis worker.
• Build cloud application logic for scan reservation, admission, job lifecycle, result state, retries, and recovery.
• Handle repeated events and uncertain network conditions through idempotency, conditional writes, immutable artifacts, and explicit state transitions.
• Integrate reconstruction and ML components into one acquisition-scoped worker with stable typed interfaces rather than a chain of network services.
• Build contract, component, integration, emulator, hardware-in-the-loop, and representative-cloud tests that exercise real failure modes, not only the expected path.
• Add structured logs, metrics, alerts, and reconciliation so that a disabled trigger, stalled scan, or duplicate execution does not go undetected.
• Collaborate with ML engineers, the cloud infrastructure lead, hardware, clinical, cybersecurity, and regulatory contributors.
• Use coding agents and engineering automation to accelerate implementation, testing, investigation, documentation, and traceability work, reviewing generated changes critically and verifying behavior before accepting them.
• Take on adjacent engineering work when it is the clearest path to completing a product feature.

Required Qualifications

• Substantial professional experience building and operating production software systems, with clear ownership of the behavior, reliability, and delivery of those systems.
• Strong Python skills and solid software-design, API, data-contract, and debugging judgment.
• Experience on both sides of at least one meaningful system boundary: device and cloud, service and infrastructure, backend and hardware, or data plane and control plane.
• Practical understanding of idempotency, retries, state machines, event-driven processing, object storage, concurrency, failure recovery, and observability.
• Strong testing practices: unit tests, integration tests, contract tests, and CI/CD.
• Ability to work from an incompletely specified goal, define a reasonable scope, identify the decisions that need broader input, and make progress without waiting for a complete specification.
• A history of completing difficult integration work, not only designing or prototyping it.
• Comfortable using coding agents as engineering tools while retaining technical accountability for the output.
• Clear communication in English with software, ML, hardware, clinical, and product collaborators.
• Strong Senior candidates typically have five or more years of relevant professional experience; candidates with fewer years and directly applicable depth are also considered. A relevant degree is useful but not required.

Preferred Qualifications

• Google Cloud, particularly GCS, Eventarc, Firestore, Cloud Run, Cloud Run Jobs, Cloud Scheduler, Secret Manager, Artifact Registry, Cloud Logging, and IAM.
• FastAPI, OpenAPI, JSON Schema, Server-Sent Events, or Pydantic-based typed contract tooling.
• Terraform or OpenTofu, GitLab CI, Workload Identity Federation, Docker containers, and GPU jobs.
• Linux device software: NVIDIA Jetson, Raspberry Pi, USB/UVC cameras, V4L2, GPIO, process supervision, local state persistence, or store-and-forward data handling.
• Image or video acquisition, FFmpeg, FFV1, lossless data handling, scientific instruments, or GPU processing pipelines.
• Medical-device software, laboratory systems, patient-identity boundaries, IEC 62304, design controls, or evidence-based development practices.
• Result delivery interfaces such as SMB file shares, HL7v2/MLLP, FHIR, SMART Backend Services, LOINC, or UCUM.
• TypeScript or a modern web framework for a limited on-device operator interface.
• Background in robotics, IoT, industrial automation, scientific instruments, imaging systems, or healthcare software, where device software and cloud services must function together as a single product.

What Success Looks Like

• Traced and run a representative acquisition through the device-to-cloud path and made its important interfaces and failure states explicit.
• Taken durable ownership of a prioritized edge or cloud product capability with a stable contract, tested behavior, and a clear definition of done.
• Landed meaningful implementation work with contract, integration, or end-to-end coverage, not only an architecture proposal or isolated prototype.
• Exercised and improved at least one high-risk reliability path involving restart, retry, duplicate delivery, conditional publication, network interruption, or recovery.
• Added operational evidence that lets another engineer reproduce the component and diagnose its major failures from versioned artifacts, logs, and documentation.
• Established productive workload sharing and peer review with the existing platform and infrastructure owner.

Our Technical Environment

Current stack: Python 3.12, FastAPI, Pydantic, PyTorch, Detectron2, OpenCV, GCS, Eventarc, Firestore, Cloud Run, Cloud Run GPU Jobs, Cloud Scheduler, Artifact Registry, Secret Manager, Cloud Logging, Docker, Terraform/OpenTofu, GitLab CI, and a modular monorepo. We use the smallest architecture that preserves the intended product behavior and recovery model. Interfaces, versions, tests, and failure behavior are product features, not afterthoughts.

Why This Role

• A real product that covers a physical scanner, cloud analysis, and a clinical delivery workflow, not an internal abstraction layer.
• Direct responsibility across camera software, Linux device services, cloud control, GPU analysis integration, and clinical result delivery.
• Difficult reliability problems with visible product consequences: restart safety, duplicate events, network outages, and idempotent result publication.
• A small team where interfaces and engineering standards are still being established.
• The chance to build regulated-product rigor into the normal development workflow from the start.

Location & Hours

Fully remote from Poland. The team works on US Eastern Time with flexible hours; we ask for reliable overlap with core Eastern Time hours (afternoons/early evenings in Warsaw) for reviews, pairing, and clinical-collaborator meetings.

B2B contract. Flexible paid time off.

illumicell AI is an equal opportunity employer. If you require a reasonable accommodation during the application or interview process, please let us know when you apply.