Prabhat Kumar Sharma

Platform Engineer | Infrastructure Architect

Specialized in building scalable distributed systems, AI/ML infrastructure, and cloud-native platforms. 3+ years of experience architecting high-performance solutions at Qualcomm and MathWorks, processing millions of GPU performance records and automating complex engineering workflows.

3+ Years Experience
50% Efficiency Gain
1M+ Records Processed
Get In Touch View Work
Prabhat Kumar Sharma
  • Bengaluru, India
  • +91 7417330025
  • prabhat.04011998@gmail.com
  • IIT Jammu (M.Tech)
Distributed Systems Cloud Native AI/ML Ops

Technical Expertise

Comprehensive skill set spanning infrastructure, backend systems, and AI/ML platforms

Programming Languages

Python C/C++ JavaScript TypeScript Golang MATLAB Shell Scripting

Backend & Distributed Systems

FastAPI Django Node.js MySQL RabbitMQ Redis MongoDB PostgreSQL

Frontend & UI

React Next.js HTML5 CSS3 Bootstrap EJS QUnit

Infrastructure & DevOps

Docker Kubernetes Jenkins CI/CD AWS ElasticSearch Kibana Git/Gerrit

AI/ML & Data Processing

Agentic AI LangChain LangGraph RAG MCP Server NumPy Pandas Scikit-learn

GPU & Performance Engineering

GPU Benchmarking Performance Analysis ARM Linux Telemetry Regression Testing

Professional Experience

Building scalable infrastructure and automation platforms at leading tech companies

Software Engineer

Qualcomm India Private Limited

Feb 2025 – Present

Bengaluru, India

Automation Strategy, AI GPU Infrastructure

  • Architected scalable automation platforms for GPU performance validation, processing millions of benchmark records across distributed systems
  • Pioneered Agentic AI and MCP-based solutions for automated root cause analysis, reducing debugging turnaround time by 50%+
  • Built AI-driven regression tracking framework aggregating build, validation, and performance data across multiple UMD variants
  • Developed multimodal AI workflows analyzing benchmark metrics, logs, and performance charts for intelligent regression detection
  • Designed GPU benchmark execution and telemetry collection pipelines for ARM-based Linux platforms
  • Modernized legacy applications through microservices architecture migration, reducing manual release effort by 70%+
50% Faster Debugging
70% Less Manual Work
1M+ Records/Day
Tech Stack: Python, Shell Scripting, Django/FastAPI, NodeJS, React/NextJS, MongoDB, Redis, RabbitMQ, Docker, Kubernetes, Jenkins, Elasticsearch, Kibana, PostgreSQL, MCP Server, Agentic AI

Software Engineer

MathWorks India Pvt. Ltd.

2023 – 2025

Bengaluru, India

  • Led Qt version upgrade initiative, enhancing organization-wide security posture and driving cross-team compatibility updates
  • Architected comprehensive error-reporting feature for lookup table data structures using JavaScript and QUnit, achieving 95% reduction in object reload times
  • Optimized model development workflow by implementing C++ based Simulink compatibility checks for production parameters
  • Engineered advanced resize and orientation control features for Simulink blocks, significantly improving model organization and developer productivity
  • Developed ML classification model optimizing build & test engine with 92% accuracy
95% Faster Reloads
92% ML Accuracy
Tech Stack: Golang, JavaScript, C/C++, QUnit, NodeJS, ReactJS, MATLAB, Simulink

Featured Projects

Innovative solutions in AI, infrastructure automation, and distributed systems

AI-Powered Regression Analytics Platform

2025 | Qualcomm

Enterprise-grade AI-driven regression tracking framework for GPU performance validation and intelligent analysis across distributed systems.

  • Multimodal AI workflows for analyzing benchmark metrics, logs, and performance charts
  • Natural-language performance analytics with automated reporting and insights
  • Reduced debugging time by 50% through intelligent root cause analysis
  • Scalable architecture processing millions of performance records daily
Agentic AI MCP Python FastAPI

GPU Performance Automation Infrastructure

2025 | Qualcomm

Comprehensive automation platform for GPU benchmark execution, telemetry collection, and performance optimization across ARM-based Linux platforms.

  • Built scalable APIs and data pipelines processing millions of performance records
  • AI-powered dashboards with real-time visualization and optimization recommendations
  • Automated CI/CD pipelines reducing manual effort by 70%
  • Microservices architecture with Docker and Kubernetes orchestration
Docker Kubernetes Jenkins MongoDB

Underwater Object Detection System

2023 | Master's Thesis

State-of-the-art underwater object detection system leveraging transformer-based few-shot learning for low-visibility environments.

  • Evaluated YOLOv6, Faster R-CNN, and DETR on real underwater datasets
  • Designed transformer-based few-shot learning approach with attention mechanisms
  • Achieved 12% accuracy improvement over baseline models
  • Published research findings and open-sourced implementation
Computer Vision Deep Learning Transformers PyTorch

Education

Strong academic foundation from premier institutions

M.Tech. (Computer Technology)

Indian Institute of Technology (IIT), Jammu

Specialized in AI/ML and Computer Vision

8.60 / 10

2020 - 2023

B.Tech. (Computer Science)

NSUT East Campus (Formerly AIACTR), New Delhi

Focus on Software Engineering and Algorithms

8.27 / 10

2016 - 2020

Get In Touch

Let's discuss how I can contribute to your team's success

Email

prabhat.04011998@gmail.com

Phone

+91 7417330025

Location

Bengaluru, India

Generated by QGenie Design Studio