Headline

Full-Stack Engineer | React · Next.js · Node.js | LLM Integration & AI Microservices | MCA @ KiiT

About

I build full-stack applications and AI-powered systems — from the React frontend to the LLM microservice running behind it. Recently, I designed and shipped the Media Analyst AI Agent: a production-style distributed system that ingests long-form audio/video, runs transcription via OpenAI Whisper, and generates structured insights using LLMs. I built every layer independently — frontend UI, API gateway, async task queues, AI microservices, and deployment pipeline. What I work with: React · Next.js · Node.js · Express · JavaScript · TypeScript · REST APIs · LLM APIs (OpenAI, Anthropic) · Whisper · Microservices · Async Queues · Docker · MongoDB What I've shipped: → Media Analyst AI Agent — end-to-end AI pipeline: audio/video → Whisper transcription → LLM-structured insights, built as a full microservices architecture → Gamma-style AI App — agentic workflow app with real-time LLM analysis, built at Google Kyoto West (Google Labs Build & Blog Marathon) → 50+ structured test cases and 30+ JIRA-tracked bugs documented at Besant Technologies What makes me different: I come from a Physics background (BSc) before completing my MCA — which means I approach systems analytically, not just syntactically. I don't glue APIs together; I think about data flow, failure modes, and scalability from the start. Graduated my MCA at KiiT (graduated 2025). Looking for SDE-1, Frontend Engineer, or AI Engineer roles at companies building real, hard technical products. GitHub: github.com/Pronaydeep

Location

Bengaluru

AI stack

antigravity

claude code

claude

Blackbox AI

ChatGPT

OpenRouter

Work experience

Full-Stack AI Developer — Gamma-style AI Application at Google Labs Build & Blog Marathon, Sep 2025 - Oct 2025 · 2 mos, Bengaluru, Karnataka, India, Built a Gamma-style AI-powered presentation and content generation application with agentic workflows as part of Google's Build & Blog Marathon event. Designed an agentic workflow layer where users input a topic and the system autonomously plans, drafts, and structures content — following a task decomposition → per-step execution → result assembly pattern Implemented real-time AI output using streaming LLM responses, delivering content progressively and reducing perceived latency significantly versus waiting for full generation Built and documented the project end-to-end within the event's time constraints, producing a working demo and technical writeup Explored agent orchestration patterns (planning, tool use, structured output) used in production AI systems at scale.

Full-Stack & AI Engineer — Media Analyst AI Agent at Company, Aug 2025 - Sep 2025 · 2 mos, Bengaluru, Karnataka, India, Designed and built a production-style distributed AI system entirely from scratch — frontend, API gateway, async task queues, and 3 independent AI microservices running in parallel. Implemented an audio/video transcription pipeline using OpenAI Whisper, processing long-form media files (60+ minutes) into time-stamped transcripts with structured JSON output Integrated LLM APIs to generate structured insights (summaries, key topics, sentiment) from raw transcripts — reducing manual review time for long-form content from hours to under 2 minutes Built an async queue system (Bull/Redis pattern) to handle concurrent media processing without blocking the API gateway, enabling the system to handle multiple jobs simultaneously Designed the microservices layer for extensibility — new AI services can be added without modifying the core gateway, following a clean separation of concerns Architected and deployed the full system end-to-end: UI → gateway → queue → microservices → output delivery

Software Tester intern at Besant Technologies, Mar 2022 - Jul 2022 · 5 mos, Bengaluru, Karnataka, India, Created and executed 50+ manual test cases Documented 30+ bugs in JIRA Performed regression, smoke, and UI testing Worked with functional specifications and user stories

Software Test Engineer at Besant Technologies, Mar 2022 - Jul 2022 · 5 mos, Drove defect detection rate of 60%+ by designing and executing 50+ manual test cases across regression, smoke, and UI suites — reducing client-reported bugs by an estimated 40% pre-release. Reduced bug reproduction time by ~30% by directly analysing backend API behaviour and functional specs with developers, cutting back-and-forth cycles on 15+ defect tickets tracked in JIRA. Owned end-to-end QA for 3 client-facing features, including UI stability checks across 5 browsers — achieving zero P0 regressions in final UAT.

Education

Master of Computer Applications - MCA, Computer application, KIIT - Kalinga Institute of Industrial Technology

Bachelor of Science - BS, physics, Dibrugarh university

Public links

github: https://github.com/Pronaydeep

leetcode: https://leetcode.com/Pronaydeep_Paul1999