Oct 2024 – May 2025 · 7 mos · Plano, Texas, United States · Remote

• Building Meshantra, a Data Mesh platform that empowers domain-specialized Data Suppliers, Builders, and Consumers to collaborate in Data Value Chains and deliver Data Products as a Service (DPaaS). The platform promotes experiential data valuation, enabling users to discover, build, and share trusted data products efficiently.
• Designed and implemented a scalable data profiling engine that extracts deep insights from structured datasets across domains. The module supports:
• 1. Automatic type inference using custom logic and the Visions library
• 2. Statistical profiling (mean, std, quantiles, missingness, cardinality, etc.)
• 3. Anomaly detection using thresholds and drift and evolution based metrics
• Benchmarking & Performance Optimization
• 1. Benchmarked profiling pipelines on large-scale datasets using Polars and BigQuery/Postgres
• 2. Optimized memory and runtime for multi-million row datasets by minimizing data scans and redundant computation

May 2024 – Aug 2024 · 3 mos · Houston, Texas, United States · On-site

• Developed a Generative AI-powered application to automate the creation, validation, and deployment of Terraform configurations for the Cloud Team at TGS. Leveraged GCP’s Vertex AI, Azure DevOps, and JIRA REST APIs to streamline infrastructure provisioning, enhancing efficiency and accuracy in infrastructure management.
• Developed a RAG-based system with a centralized vector store using Postgres to manage text embeddings of Terraform deployments across teams, enabling efficient search and retrieval of release history for improved operational insights.

Aug 2023 – May 2024 · 9 mos · Arizona State University, Tempe · On-site

• Developed SEVD, a first-of-its-kind multi-view synthetic event-based dataset using the CARLA simulator, designed to support research in autonomous driving and event-based perception.
• Motivation
• Conventional RGB cameras often face limitations in handling challenging dynamic conditions such as poor lighting or adverse weather. Event-based vision sensors provide a promising alternative, but publicly available datasets remain limited. SEVD addresses this gap by providing a comprehensive and diverse synthetic dataset.
• Key Highlights
• 1. Captured using multiple dynamic vision sensors in both ego and fixed-view setups.
• 2. Diverse environmental conditions:
• Lighting: Noon, twilight, and nighttime
• Weather: Clear, cloudy, wet, rainy, and foggy
• Domain shifts: Includes both discrete and continuous variations
• 3.Covers a wide range of scene types: urban, suburban, rural, and highway
• 4. Includes multiple object classes: cars, trucks, vans, bicycles, motorcycles, and pedestrians
• Multimodal Data Included
• Alongside event data, SEVD provides: RGB imagery, Depth maps, Optical flow, Semantic segmentation, Instance segmentation
• Evaluation and Benchmarks
• Evaluated with both event-based models (e.g., RED, RVT) and frame-based models (e.g., YOLOv8) for traffic participant detection
• Generalization Experiments
• Performed studies to assess the generalization ability of synthetic event-based data across different conditions and tasks.
• Availability
• The dataset is publicly available at: https://eventbasedvision.github.io/SEVD

Apr 2022 – Jul 2023 · 1 yr 3 mos · Bengaluru, Karnataka, India

• As a Software Development Engineer - II at BrightCHAMPS, I played a pivotal role in architecting and building robust cloud-based communication solutions. Leveraged AWS Cloud services, including Kinesis Streams, Lambda Functions, API Gateway, and Step Functions, to achieve a 30% reduction in communication failures across the company. Adept in designing load-balanced, scalable web applications using EC2, Lambda, API Gateway, RDS, and Elastic Cache (Redis), applying micro-services and serverless design patterns. Managed a cross-functional team, fostering seamless collaboration, and achieving project milestones.

May 2021 – Apr 2022 · 11 mos · Bengaluru, Karnataka, India

• During my tenure as a Software Development Engineer I at Brightchamps, I contributed to design and development of a comprehensive system to efficiently match students with the most suitable teachers based on parameters such as grade level, language proficiency, and historical performance records.

Aug 2020 – May 2021 · 9 mos · Bengaluru, Karnataka, India

• As an Intern at Brightchamps, I actively contributed to the development of REST API endpoints, which served as the backbone for efficient communication within the system. This involved designing, implementing, and testing API functionalities to ensure seamless data exchange and interaction between different components of the application.
• In addition, I played a significant role in creating and enhancing various dashboards tailored for monitoring system performance and user activity. These dashboards provided valuable insights into key metrics, allowing the team to make informed decisions and identify areas for optimization. Leveraging visualization tools and best practices, I designed intuitive user interfaces that facilitated real-time monitoring and analysis.
• Furthermore, I collaborated closely with senior engineers to troubleshoot issues, improve code quality, and enhance overall system functionality.

May 2019 – Jul 2019 · 2 mos · Hyderabad, Telangana, India

• IV therapy is an easy and effective procedure which allows a drop by drop administration of medication. The process also has some innate limitations, like formation of tissue in the needle, rolling of patient on the tube or over the hand, can block the flow of the fluid and compromises with the patient’s life in severe scenarios. The major bottleneck of the process is monitoring the medicine bottle.
• Developed an automated intravenous fluid monitoring system using infrared drop detection and non-contact liquid level sensing technologies, enhancing patient safety by reducing manual monitoring and the risk of critical complications.
• Leveraged Arduino Uno R3 to integrate drop counting via laser-light dependent resistor and real-time liquid level assessment with a non-contact sensor, showcasing innovative healthcare technology and decreasing the need for constant nurse supervision.