Website:
dusq.com
Job details:
About DUSQ
DUSQ builds a behind-the-ear wearable combining PPG and motion sensing with transcutaneous auricular vagus nerve stimulation (taVNS) to improve sleep and autonomic health. Users wake up to health scores and coaching built from their own physiological data.
Work hours: 8 hrs/day, 11 AM- 7 PM
Location: In-office, Dwarka, New Delhi
Expected Salary: 25-30 LPA
About the Role
Own your own data pipelines and architecture from 0 to 1; the full path from raw device and app data to the interactive metrics and insights users see. The role combines machine learning, signal processing, and backend engineering, deployed into a live application serving real users.
Key Responsibilities
● Strong command of machine learning and deep learning; model evaluation, regularization, ensemble methods, and neural architectures for time-series data (CNNs, LSTMs, Transformers), with lifecycle ownership of production models: zero-downtime migrations, accuracy and latency optimization, and model-behavior visualization (PyTorch, scikit-learn, MLflow, Plotly).
● Proven experience in digital signal processing; filter design, time/frequency-domain analysis, spectral estimation, artifact rejection, applied to real-world data (NumPy, SciPy).
● Architect and deploy the app's personalization engine; fusing user scores, multi-night context, in-app behavior, and user interactions to drive the content, insights, and notifications that populate the app.
● Working proficiency with backend infrastructure: relational and time-series databases (MongoDB, Redis), job scheduling and orchestration (Celery, Airflow), monitoring and cloud deployment (AWS, Docker).
● Partner with R&D, product, and engineering on rapid prototyping, hypothesis validation across user segments, and controlled releases to a live application (Git, CI/CD).
● Engineer lightweight on-device inference models (TinyML); compact 1D CNNs and regression models within strict memory and compute budgets, to improve classical signal-processing stages (TensorFlow Lite, ONNX Runtime, quantization/pruning).
● Build internal tooling and admin dashboards that standardize user data into decision-ready views for operations and coaching teams (Streamlit/Dash, FastAPI).
Qualifications
● Degree in Data Science, Computer Science, Electrical/ECE/Bio-medial Engineering, or a related field.
● Relevant hands-on experience in data science or applied research.
● Research experience; academic or industry is a plus.
● What we weigh most: the systems you have built and what they delivered.
Click on Apply to know more.