Udaya Kiran Garudachalam

Project: Qualcomm - ADAS (Advanced Driver Assistance Systems) Jun 2025 – Present
Role: Devloper
Tech Stack: Python, SQL, pandas, PyArrow, AWS S3, boto3, AWS CLI, Docker, GitLab CI/CD, JFrog, DataHive, SIMPHERA, Grafana, Foxglove Studio, Pintor

• Automated end-to-end ADAS re-simulation workflows in DataHive: ingestion, sequencing, syncing, re-simulation execution, configuration updates, conflict resolution, and result validation.
• Implemented automation scripts to handle high-volume datasets efficiently, improving throughput, and reducing manual intervention.
• Implemented MF4 synchronization using AWS S3 (DataHive), Python, boto3, and AWS CLI to ensure dataset consistency across simulation pipelines.
  Containerized ingestion, MF4 validation, event extraction, and metadata enrichment using Docker, improving reproducibility and performance.
• Automated CI/CD via GitLab for pipeline execution, job scheduling, and deployment of simulation automation scripts.
• Designed CLI utilities for dataset browsing and filtering (session, date, hashkey), selective MF4 downloads, and VCI retrieval from S3—reducing data retrieval time from hours to minutes.
• Developed SQL routines to fetch and validate data required for pipeline execution and reporting.
• Architected an end-to-end validation suite for release outputs (Parquet) using PyArrow/Pandas with schema and logic checks, Matplotlib 2D diagnostics, and consolidated HTML reports, ensuring 100% data accuracy prior to deployment.
• Created multiple daily process automation scripts to significantly reduce execution time and improve operational reliability.
  

Project: NextGen R&D (TCS internal) Aug 2023 – Jun 2025

Role: Developer
Tech Stack: Python, Linux, open-source tools, LLM workflows, RL automation

Project: Game Automation Powered by LLM Intelligence

• Automated gameplay using RL-based logic and extracted detailed telemetry and environment signals (including hardware-related anomalies).
• Designed a pipeline to feed game descriptions, telemetry into an LLM for test case generation, and script generation.
• Supported the execution and validation of generated scripts within the game environment to address the described scenarios.

Project: Game Testing (Daisy Chaining Approach)

• Implemented a daisy-chaining workflow by integrating multiple open-source tools and applying them on a game to generate a consolidated report.
• Evaluated tools, captured output metrics, and integrated the toolchain to produce end-to-end consolidated results.

Project: Linux Device Driver Testcase Generation using Open-Source LLM

• Developed Linux device driver test cases in Python to validate driver performance and behavior.
• Generated test cases using Gen-AI approaches, and implemented validation code to verify the authenticity and correctness of AI-generated test cases.