Nawin Kumar Sharma

Address: Bengaluru, India

Phone: +91 9835979046

LinkedIn: linkedin.com/in/nawin-sharma-83009825a

<aside>

Software Engineer | Full Stack & AI

Builder. Problem-solver. Full stack dev. Always exploring the edge of what’s possible with tech—shipping scalable, intuitive products that solve real-world problems. Lately: AI and real-time systems.

</aside>

Professional Experience

<aside>

Software Development Engineer Intern — Lumen Technologies | Feb 2026 – Aug 2026

Rebuilt 2 business-critical internal tools, migrating from legacy ASP.NET to Java/Spring Boot, owning the full stack (frontend, backend, database, testing, and deployment).
Helped reduce operational costs by an estimated $60K+ annually and cut page load times from ~2.5s to ~300ms by replacing maintenance-heavy legacy systems with optimized, scalable services.
Shipped via Jenkins CI/CD pipelines on Rancher-managed Kubernetes clusters on Azure, backed by JUnit unit and end-to-end tests.

Bengaluru, India

</aside>

<aside>

Full Stack Developer — CarboInsights | Jul 2025 – Feb 2026

Built a multi-tenant B2B SaaS platform (React, TypeScript) adopted by 5 manufacturing companies, processing 1M+ emission records with real-time dashboards and automated Excel reporting.
Designed a configurable Multi-Framework Emission Engine supporting 5+ regulatory frameworks (WSA, CBAM, Green Steel) via dynamic SQL generation and variable emission factors—cutting compliance reporting from hours to minutes.
Engineered an intelligent AI chatbot using the AI SDK and RAG, combining retrieval with live database insights and persistent, context-aware sessions for emission analysis and sustainability recommendations.

Nashik, India (Remote)

</aside>

Key Projects

<aside>

StreamAI | 2025

Built a production-grade AI chat platform with streaming responses and virtualized rendering, sustaining sub-second perceived latency across long multi-turn conversations.
Implemented RAG over PDFs, websites, raw text, and YouTube transcripts, indexed in Qdrant for fast semantic search and grounded answers.
Integrated long-term memory via Mem0 to persist user facts, preferences, and context across sessions for personalized responses.

Stack: Next.js, AI SDK, Qdrant, LangChain, Prisma

Website: https://streamai.nawin.xyz/

GitHub: https://github.com/nawinsharma/streamai

</aside>

<aside>

WatchNChill | 2026

Built a real-time YouTube watch party platform with synchronized playback, per-room live chat, and role-based host controls (full-duplex WebSockets via Socket.IO).
Engineered a custom useVideoSync hook bridging the YouTube IFrame API with Socket.IO to broadcast play/pause/seek with periodic re-sync, maintaining ~200ms drift during network fluctuations.
Reduced backend database traffic by persisting room/session state in Redis, enabling stateless WebSocket servers that scale horizontally.
Implemented host promotion + auto-close on host exit; validated client payloads with Zod for type-safe, conflict-free playback control.

Stack: Next.js, Socket.IO, Redis, YouTube API

Website: https://watchnchill.nawin.xyz/

GitHub: https://github.com/nawinsharma/WatchNChill

</aside>

Technical Skills

<aside>

Languages: JavaScript, TypeScript, Java, Python
Frontend: React, Next.js, Tailwind CSS, shadcn/ui, React Native (Expo)
Backend: Node.js, Express.js, Spring Boot, Socket.IO, Redis, RESTful APIs, WebSockets, JUnit
Databases: PostgreSQL, MongoDB, SQL, Prisma, Drizzle
AI / ML: Generative AI, RAG, Tool Calling, AI SDK, LangChain, Qdrant, Mem0, LLM APIs
DevOps & Cloud: Git, Docker, Kubernetes, Jenkins, CI/CD, AWS (EC2, S3), Azure, Rancher, Turborepo
Concepts: Distributed Systems, System Design, Microservices, Real-Time Systems, Agile/Scrum </aside>

Education

Jain (Deemed-to-be University), Global Campus — B.Tech, Computer Science and Engineering (CGPA: 8.6/10)

Bengaluru, India — 2022 – 2026

</aside>

Trinity International College — Higher Secondary Education (Science — Computer) (GPA: 3.68/4.0)

Kathmandu, Nepal — 2020 – 2022

</aside>

Publications

<aside>

Research Publication — Comparative Study of Code Generation by AI Models

IEEE, ICCRTEE 2025 — May 2025

</aside>