The Role : We are looking for a Machine Learning Engineer to build and productionize models that power fall detection, vitals monitoring, and predictive health insights from radar sensor data.

Location : Bangalore

Work Mode : Work from Office

You will work closely with hardware, data engineering, backend, and product teams to improve model accuracy, reduce false alarms, and deploy reliable ML systems into production.

This role is ideal for someone who is strong in classical ML, comfortable with messy real-world sensor data, and able to write clean production-grade code.

What You'll Do :

- Build classical ML models such as XGBoost, ensembles, anomaly detection, and time-series methods for fall detection, vitals monitoring, and health risk scoring.

- Engineer features from raw, sparse, and noisy radar signal data, point-cloud data, and time-series sensor streams.

- Contribute to CV-adjacent work such as pose, skeleton, movement, and activity estimation from radar data.

- Build data pipelines on Databricks for training, evaluation, and inference workflows.

- Perform exploratory data analysis on resident, device, alert, and facility-level data to identify patterns, edge cases, and model improvement opportunities.

- Own model evaluation for a safety-critical system, including precision, recall, sensitivity, specificity, false alarms, missed events, and detection latency.

- Analyze production model behavior across facilities, residents, devices, and time periods.

- Work with noisy real-world data, including missing values, label quality issues, device variation, sparse events, and facility-specific patterns.

- Write clean, modular, tested Python code for ML training, evaluation, feature engineering, and inference.

- Deploy, monitor, and improve models in production.

- Work closely with hardware and data engineering teams to improve data quality, labeling, observability, and model reliability.

What We're Looking For :

- 3-4 years of experience building and shipping ML systems in production.

- Strong Python programming skills, with the ability to write maintainable, testable, production-grade code.

- Strong classical ML fundamentals : feature engineering, model training, cross-validation, error analysis, and model evaluation.

- Experience with models such as XGBoost, Random Forests, gradient boosting, ensembles, anomaly detection, or time-series models.

- Good SQL skills and ability to analyze large datasets using SQL, PySpark, pandas, or Databricks.

- Experience working with time-series, sensor, spatial, point-cloud, IoT, or computer vision-like data.

- Working knowledge of data engineering workflows, preferably using Databricks, Spark, Delta Lake, or similar platforms.

- Strong debugging and analytical skills : able to trace issues across data, model, pipeline, and production behavior.

- Comfortable working with ambiguity in a small, fast-moving team.

- Strong ownership mindset and ability to take a model from research/experimentation to production.

Good to Have :

- Experience in healthtech, IoT, radar, wearables, ambient monitoring, or safety-critical systems.

- Exposure to computer vision, pose estimation, skeleton tracking, object tracking, or spatial data.

- Experience with MLflow, model registry, feature stores, model monitoring, or experiment tracking.

- Experience with ONNX, quantization, edge deployment, latency optimization, or resource-constrained inference.

- Familiarity with streaming data pipelines, Kafka, Spark Structured Streaming, or real-time inference systems.

Key Responsibilities :

- Production ML model delivery

- Cross-functional ML collaboration

- Radar-based activity algorithms

- Predictive health insights development

- Radar data feature engineering

Requirements :

MUST HAVE :

- 3-4 years of experience