2 days ago
Bengaluru, IndiaSenior
Responsibilities
- Design, develop, train, fine-tune, validate, and optimize machine learning and deep learning models for ADAS and automotive validation use cases.
- Build end-to-end ML pipelines for data acquisition, dataset curation, labeling, preprocessing, cleaning, feature engineering, model training, evaluation, and experiment tracking.
- Develop DNN, CNN, RNN, Transformer, and other state-of-the-art models for perception, signal processing, event detection, and validation workflows.
- Define training, validation, and test datasets and establish data quality standards, including analysis of class imbalance, labeling accuracy, feature distributions, and data drift.
- Implement distributed model-training workflows on AWS, Azure, and HPC environments and perform hyperparameter tuning, benchmarking, and error analysis.
- Develop evaluation frameworks and validation methodologies and investigate model failures with algorithm and validation teams.
- Create automated processing pipelines for camera, radar, CAN, Ethernet, vehicle telemetry, and other automotive sensor data.
- Participate in vehicle testing, instrumentation, logging setup, data collection, scenario identification, and data validation.
- Support deployment-readiness reviews, document model performance and limitations, and contribute to ADAS V-cycle processes and continuous improvement of data-analysis tools.
Requirements
- Bachelor's or Master's degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, Electrical Engineering, Computer Engineering, or a related discipline.
- 6-8 years of industry experience in Machine Learning, Data Science, Deep Learning, or related fields.
- Strong Python development skills and experience building production-grade ML pipelines.
- Demonstrated experience training machine learning and deep learning models from scratch using large-scale datasets.
- Hands-on experience with dataset creation, collection, labeling, cleaning, augmentation, feature engineering, and data quality assessment.
- Strong understanding of statistical learning, model evaluation metrics, validation methodologies, and error analysis.
- Experience with large-scale structured and unstructured datasets.
- Understanding of automotive signals, ECUs, CAN, Ethernet, and embedded-systems fundamentals.
- Preferred experience with ADAS, autonomous driving, robotics, automotive perception datasets, camera, radar, lidar, and vehicle sensor data.
- Preferred experience with Transformer-based and foundation models, MLOps tools, model explainability, bias analysis, robustness testing, GPU optimization, distributed training, and large-scale data processing.
- Exposure to C/C++, MATLAB, signal processing, embedded software, CANoe, CANalyzer, SIL/HIL environments, and automotive validation workflows.
Benefits
- Higher education and continuous-development opportunities through Udacity, Udemy, and Coursera.
- Life and accident insurance.
- Sodexo cards for food and beverages.
- Well Being Program with workshops and networking events.
- Employee Assistance Program.
- Access to fitness clubs, subject to terms and conditions.
- Creche facility for working parents.
- Global, inclusive workplace with support for physical and mental health.
Categories
About Aptiv
Aptiv is a public automotive technology supplier that designs and manufactures vehicle electrical architectures, ADAS/autonomous software, connectivity solutions, and engineered components. It sells integrated hardware-software platforms, wiring and power distribution systems, sensors, and infotainment to global automakers and mobility companies. Headquartered in Dublin, Ireland, Aptiv is listed on the NYSE and operates worldwide across development centers and manufacturing sites.