Shubham Shrivastav - Senior Full Stack Software Engineer

Professional Summary

Senior Full Stack Software Engineer with 9+ years building high-craft web applications and AI-powered platforms. Expert in Angular, React, TypeScript, Node.js, Python, PostgreSQL with deep experience in LLM integration, AI orchestration, RAG, and evaluation systems. Product-minded engineer who architects complex features end-to-end.
Built NexDocs (AI publishing platform with multi-provider LLM orchestration, RAG pipeline, and evaluation framework) and AutoResearch (multi-agent AI research system using CrewAI). Mentored engineers, champion technical excellence, and drive measurable impact through scalable architecture. Selected as top 300 builder in India (out of 25,000+ applicants) for Vibecon.com. Remote since 2018.

Technical Skills

AI & LLMs: LLM Integration (OpenAI, Anthropic, Gemini) • RAG • pgvector • Prompt Engineering • AI Orchestration • Evaluation Frameworks • Embeddings • SSE Streaming

Frontend: React • Angular • TypeScript • JavaScript • Next.js • HTML • CSS • Tailwind CSS

Backend: Node.js • NestJS • Python • FastAPI • GraphQL • REST APIs • Microservices • Firebase Functions

Data & Infrastructure: PostgreSQL • Redis • MongoDB • Prisma • pgvector • Docker • AWS • System Architecture

Tools: Git • CI/CD • GitHub Actions • Jest • Cypress • A/B Testing • NX Monorepo • Code Reviews

Professional Experience

Full Stack Software Engineer

Luzmo (Formerly Cumul.io)

January 2018 – Present Remote

JavaScript, TypeScript, React, Angular, Node.js, GraphQL, Python, FastAPI, PostgreSQL, Redis, MongoDB, Docker

AI-powered product development: Partner closely with Product and Design to architect and ship complex features including LLM-powered capabilities for analytics SaaS serving thousands of concurrent users. Delivered 40% performance improvement in AI features through Redis caching, query optimization, and async workflows.
Complex feature architecture: Led technical design and implementation of interactive dashboards, real-time data visualization, and content publishing tools. Built scalable REST APIs and microservices using Node.js/TypeScript with PostgreSQL, processing high-throughput requests with low latency.
Legacy codebase evolution: Architected micro-frontend migration achieving 60% faster builds while safely refactoring legacy code. Applied systems thinking to enhance scalability and maintainability, championing patterns that elevated team-wide standards.
Technical leadership: Mentored 5+ engineers through code reviews, pairing sessions, and architecture discussions. Implemented comprehensive testing (Jest, Cypress) increasing coverage from 65% → 92%, reducing production issues.

Software Engineer

Persistent Systems Ltd.

September 2016 – September 2017 Nagpur, India

Python, Django, JavaScript, HTML, CSS, React, PostgreSQL, OCR, Tesseract

Cross-functional collaboration: Worked with Product and Design teams to create user-focused solutions. Built automation tools saving 120+ IT hours per cycle.
Quality-driven execution: Developed OCR translation system eliminating errors for thousands of documents monthly. Built secure applications for government clients with thorough testing.

Projects

NexDocs - AI-Powered Publishing Platform

Next.js, React, TypeScript, NestJS, FastAPI, PostgreSQL (pgvector), Redis, OpenAI, Anthropic, Gemini

Architected AI orchestration layer with multi-provider LLM integration (OpenAI, Anthropic, Gemini) for intelligent content generation with provider fallback and cost optimization
Built RAG pipeline using pgvector with OpenAI embeddings for context-aware retrieval and generation, including chunking strategies and semantic search
Designed prompt management system and evaluation framework to measure AI output quality, reliability, and cost/latency tradeoffs
Implemented real-time LLM streaming via Server-Sent Events (SSE) with fault-tolerant circuit breakers, achieving 70% faster content generation through parallel processing

AutoResearch - AI-Powered Research Agent

Python, Flask, CrewAI, LangChain, OpenAI, Gemini, Groq, SerperDev, Flutter, Firebase, Cloud Firestore

Built multi-agent AI system using CrewAI with specialized Researcher and Reporter agents orchestrated in sequential pipelines for automated research synthesis
Implemented multi-provider LLM orchestration (OpenAI GPT-4o, Google Gemini, Groq Llama3) with environment-based model selection and tool use (web search, website RAG)
Designed production backend with Firestore-based rate limiting, Firebase App Check security, and research paper analysis via LLM-powered web scraping
Built cross-platform Flutter frontend (Android/iOS) with markdown rendering, voice input, and automated PDF report generation

Ocularis - Autonomous Browsing Agent

Python, FastAPI, Playwright, PostgreSQL, pgvector, Pydantic, Next.js, TypeScript, Docker

Built open-source autonomous web agent that converts high-level goals into actionable browsing plans with postcondition checks and structured completion criteria
Implemented reliability layer with screenshot-hash stuck detection (MD5 + SSIM), automated recovery strategies, and circuit-breaker pause for human-in-the-loop escalation
Designed context-aware extraction and reasoning flow to identify relevant candidates from noisy pages, compare options, and return concise structured answers

TimeDial - macOS Timezone App

SwiftUI, AppKit, Combine, macOS

Built native macOS menu bar app for remote workers managing multiple timezones with seamless system integration
Implemented interactive analog clocks with real-time sync and drag-to-adjust time across all clocks
Integrated smart timezone search with 450+ zones and published to App Store with focus on performance

Education

Bachelor of Engineering

L.N.C.T, Bhopal

Graduated 2016