What I've built
Each role shaped how I think about data systems, team leadership, and delivering at scale.
CaliperAI
Co-founder
The Challenge
ML teams needed a unified platform to annotate images, 3D point clouds, and medical imaging data—existing tools were fragmented and couldn't scale.
What I Built
Architected a high-performance annotation platform with multi-modal support. Built a RAG-based search assistant for complex event retrieval from automotive time-series datasets.
End-to-end ownership from architecture to production, enabling rapid iteration on ML training workflows.
Ielektron
Solutions Architect
The Challenge
Multiple automotive software projects had inconsistent DevOps practices, slowing down delivery and making cloud modernization difficult.
What I Built
Designed and implemented unified CI/CD pipelines across projects. Led cloud and data modernization initiatives aligned with business requirements.
Bridged engineering teams and leadership, enabling faster delivery and consistent infrastructure across the organization.
SafeAI Inc.
Senior Data Engineer
The Challenge
Autonomous vehicle teams needed processed camera and LiDAR data at scale for ML training, but lacked the pipeline infrastructure.
What I Built
Designed annotation and ETL pipelines for large-scale sensor datasets. Built a data framework to prepare processed data for model training, plus dashboards for tracking pipeline health.
Accelerated the ML team's ability to iterate on perception models with high-quality, processed training data.
Mercedes-Benz R&D India
Senior Technical Lead
The Challenge
Enterprise data solutions across multiple projects needed coordination, and teams lacked tooling for ADAS event analysis.
What I Built
Led a 10-member cross-functional team. Architected ETL pipelines using Databricks, Airflow, and Azure DevOps. Built a web-based ADAS event analysis tool (FastAPI + React).
Managed 6 concurrent data engineering projects including mission-critical systems used across business units.
Mercedes-Benz AG
Expat Engineer (Germany)
The Challenge
Offshore collaboration between Germany and India lacked established processes for data engineering work.
What I Built
Established the collaboration model and centralized processes for data collection requirements. Automated Python library generation for backend customer data decoding.
Created the foundation for sustainable cross-border data engineering partnerships.
Mercedes-Benz R&D India
Senior Engineer
The Challenge
Raw ADTF vehicle data needed to be transformed into analysis-ready formats at scale, and test automation was manual and slow.
What I Built
Designed high-throughput ETL pipelines processing ~1TB/day, transforming raw data into Parquet. Improved pipeline performance by 3x using parallel processing.
Reduced test development effort by ~40% through automation. Participated in vehicle testing in Spain and acquired FB2 Driving Authorization.