Open to opportunities

Danya Balakumaran

@danyabalakumaran

Deep learning researcher building real-time autonomous driving and multimodal AI systems with PyTorch and deployment rigor.

India

Message

What I'm looking for

I’m looking for a team where deep learning meets production: real-time systems, computer vision or multimodal AI, strong engineering reviews, and clear ownership from data to deployment—so my models perform reliably in real-world conditions.

I’m a deep learning research-focused engineer, currently driving end-to-end autonomous driving work in CARLA—building a pipeline from 20K+ RGB frames to a CNN that achieves 95% lane-keeping accuracy and real-time 30 FPS inference.

I care deeply about reliability under real conditions, so I’ve combined CNN inference with adaptive classical feedback, using confidence-aware arbitration to maintain lane accuracy within 0.25m even under distribution shift. I also design scalable training pipelines (mixed-precision CUDA AMP, class-balanced sampling, and automated evaluation tracking) so performance and failure modes are measurable.

Beyond computer vision, I’ve built ETL and analytics pipelines (100K+ household energy records) with feature engineering and clustering, and I’ve applied optimization with Gurobi for interpretable explanations. Earlier, I shipped backend systems with Spring MVC and OracleDB to AWS RDS, established CI/CD with Jenkins (reducing deployment failures by 60%), and improved frontend performance—so I bring both research depth and production instincts to every model I deploy.

Experience

Work history, roles, and key accomplishments

Autonomous Driving Research Assistant

University of Virginia

Feb 2025 - Mar 2026 (1 year 1 month)

Built an end-to-end autonomous driving system in CARLA using 20K+ RGB frames, training a CNN in PyTorch to achieve 95% lane-keeping accuracy and deploying real-time inference at 30 FPS. Developed confidence-aware adaptive control blending and a CUDA AMP training pipeline with automated evaluation of collision rate and trajectory error.

CARLA PyTorch Computer Vision CNN CUDA Mixed Data Sampling Linux

Energy Data Research Assistant

UVA Biocomplexity Institute

Jan 2024 - Aug 2024 (7 months)

Engineered an ETL pipeline for 100K+ household energy records, applying feature engineering and K-Means clustering to identify consumption patterns across 12 demographic attributes. Used Integer Linear Programming in Gurobi to optimize interpretable explanation descriptor size and built Matplotlib visualization dashboards.

Python ETL Pandas NumPy scikit learn Gurobi Data Visualization Matplotlib Integer Programming

Associate System Engineer

Tata Consultancy Services

Feb 2022 - Jun 2023 (1 year 4 months)

Led deployment of a monolithic backend with OracleDB on AWS RDS and designed microservices to enable zero-downtime releases. Implemented scalable REST APIs and established a Jenkins CI/CD pipeline with security scanning and automated tests, reducing deployment failures by 60% for daily production releases.

Java Spring MVC REST APIs Microservices AWS RDS Jenkins CI CD Git Jira

Front-end Intern

Taigha India Private Limited

Apr 2021 - Dec 2021 (8 months)

Implemented UI enhancements and improved frontend performance by troubleshooting issues and applying lazy loading with a cross-functional team. Achieved a 35% improvement in platform responsiveness.

HTML CSS JavaScript PHP Shell Scripting Lazy Loading GitHub

Education

Degrees, certifications, and relevant coursework

University of Virginia

Master of Engineering (M.Eng), Computer Engineering

2023 - 2024

Grade: 3.6/4.0

Master of Engineering (M.Eng) in Computer Engineering at the University of Virginia (GPA 3.6/4.0) from Aug 2023 to Dec 2024.

PSG Institute of Technology and Applied Research

Bachelor of Engineering (B.E), Electronics and Communications Engineering

2017 - 2021

Grade: 8.4/10.0

Bachelor of Engineering in Electronics and Communications Engineering at PSG Institute of Technology and Applied Research (GPA 8.4/10.0) from Aug 2017 to Apr 2021.