Junior Algorithm Engineer

DUSQ (formerly InnerGize) · Full-time · [City], India (On-site) · 0–3 years experience

You're the execution engine of the algorithm team — turning senior specs into clean, tested code, building the tooling that lets the team move fast, and developing deep biosignal intuition by handling more real physiological data in six months than most engineers see in years. This role is a deliberate ramp: strong performers grow into owning algorithms end to end.

What you'll do

Implementation

• Implement PPG / HRV / recovery / sleep algorithm specifications in production-quality Python (and progressively C), with unit tests and documented edge-case behavior.
• Contribute to the Signal → Features → Scoring pipeline under senior direction.

Tooling & Data

• Build and maintain data tooling: parsers for device binary and BLE stream formats, dataset labeling pipelines, and interactive signal-visualization dashboards (HTML / Plotly).
• Run the replay / regression harness — re-execute algorithm versions against recorded nights, compute agreement metrics against reference hypnograms, and produce regression reports for every proposed change.

Validation Support

• Assist in lab + free-living data collection and quality control — flagging sensor dropouts, timestamp anomalies, and labeling inconsistencies before they poison a dataset.
• Reproduce and characterize reported algorithm failures from raw data; write clear findings with plots and numbers.

Learn the Domain

• Sleep physiology, PPG signal characteristics, HRV, SQI concepts, and the statistics under our detection logic — you'll read papers, not just tickets.

Must have

• 0–2 years of experience. Strong fundamentals in Python or MATLAB — numerical computing, data handling, plotting (NumPy / pandas / Plotly, or MATLAB Signal Processing Toolbox) — plus solid data structures & algorithms. Production code is Python; strong MATLAB engineers ready to ship in Python are very welcome.
• Basic signal processing (sampling, filtering, FFT) from coursework or demonstrated self-study — you know what aliasing does to your data.
• Careful, skeptical habits: check units, plot before you trust, and say "I don't know" instead of guessing.
• Comfort with version control (Git), code review, and readable tested code.
• Genuine curiosity about physiology and sleep science.

Nice to have

• Any exposure to C / embedded programming.
• Projects involving time-series, wearables, or biosignals (personal projects count).
• Basic scikit-learn / ML workflow familiarity.

Competitive compensation + meaningful early-stage equity. You'll see your work running on real hardware, on real people, within weeks — not quarters.

To apply, share your CV and any relevant shipped work or publications.