First Edition ยท 2026

Nitish Kumar
Purohit

A compendium of engineering, architecture,
and artificial intelligence.

SDE-III ยท Full Stack AI Engineer ยท GenAI & AI Systems
Noida, India ยท knitish899@gmail.com ยท +91-8210510438
Available for new chapters
Chapter I

About the Author

Over the course of six years in production software engineering, the author has developed a particular expertise in building systems that operate at scale โ€” systems designed not merely to function, but to endure under the weight of real-world concurrency, security audits, and the unrelenting demands of enterprise users.

For the past two years, this work has increasingly centered on Generative AI: specifically, the construction of Retrieval-Augmented Generation (RAG) pipelines, the design of agentic AI workflows, and the careful optimisation of vector database queries โ€” all deployed into production environments where failure is not abstract but measurable.

Note to the reader: The metrics presented throughout this document are drawn from production systems, not academic exercises. Each figure represents measurable, shipped outcomes.

The tools of choice include Next.js (App Router), advanced TypeScript, and FastAPI โ€” combined with a deep focus on security, performance, and the invisible craft of making complex systems feel simple. Currently based in Noida, India.

6+
Years shipped
2+
Years GenAI
40%
LCP improved
150ms
p99 latency
Figure 1. Key performance indicators from production systems, 2019โ€“2026.
Chapter II

Technical Proficiencies

The following table enumerates the primary technologies and methodologies employed across production systems. They are organised by domain for the reader's convenience.

Table 1. Complete technical stack, grouped by discipline.
Domain Technologies
Frontend React.js Next.js (App Router) TypeScript JavaScript Tailwind CSS shadcn/ui Radix UI Redux Jotai TanStack Query
Backend Python FastAPI Flask Node.js Express.js GraphQL WebSockets Socket.IO SSE
AI / GenAI RAG Pipelines Pinecone Weaviate Agentic Workflows Prompt Engineering LLM Inference MCP
Security & DevOps OAuth2 / JWT RBAC Zod Validation AWS Docker CI/CD GitLab
Architecture Microservices SaaS Architecture Core Web Vitals Vitest Jest
Chapter III

Professional History

What follows is a chronological account of professional roles held, with particular attention to measurable outcomes and the technical decisions that produced them.

SDE-III Jan 2022 โ€” Present
Mobcoder ยท Noida, India
  • Architected multi-tenant SaaS platform on Next.js App Router with zero-downtime CI/CD, handling high-concurrency traffic at scale.
  • Built production RAG pipelines (OpenAI + Pinecone/Weaviate); cut irrelevant AI responses by 60% via hybrid retrieval and re-ranking.
  • Shipped agentic AI workflows into customer-facing products โ€” 30% engagement uplift within 90 days.
  • Hardened all services with RBAC, JWT, and Zod validation โ€” zero critical vulnerabilities across three consecutive security audits.
  • Optimised FastAPI async workers + connection pooling: API p99 latency dropped from ~800ms to under 150ms.
  • Led frontend audit across 3 products; LCP improved 40%, Lighthouse scores 60 โ†’ 90+ via streaming SSR.
Full Stack Developer Jun 2020 โ€” Dec 2021
Siscaso
  • Delivered MERN stack SaaS apps for 10k+ MAU; migrated REST โ†’ GraphQL, reducing over-fetching and improving mobile performance.
  • Cut initial bundle size by 35% via code splitting and lazy loading, improving time-to-interactive on data-heavy dashboards.
Software Developer Aug 2019 โ€” May 2020
DreamBig Networks
  • Built and shipped full-stack web features using JavaScript, Node.js, and MySQL in a fast-paced production environment.
  • Collaborated with cross-functional teams to deliver user-facing features on tight deadlines.
Chapter IV

Selected Works

The projects described herein represent a selection of the most significant systems designed and delivered. Each entry includes the technical approach and its measurable impact.

Akari โ€” AI-Integrated Web Platform

2025 โ€“ 2026

Led full-stack development of an AI-powered platform with LLM-driven features and intelligent context-aware workflows. Owned end-to-end technical decisions โ€” system design, API architecture, and deployment โ€” delivering iterative feature releases on schedule.

Stack: Next.js FastAPI LLM Agentic AI

TIFIN IP โ€” AI Annuity Recommendation Engine

2022 โ€“ 2024

Built a RAG pipeline with Pinecone vector store for personalised annuity recommendations โ€” 45% retrieval precision improvement over keyword search. Integrated streaming LLM inference via SSE into a Next.js + FastAPI stack, cutting perceived latency and boosting advisor satisfaction.

Stack: RAG Pinecone SSE Next.js FastAPI

TIFIN Give โ€” DAF Philanthropy Platform

2022 โ€“ 2023

Built multi-custody grant management and community giving flows in Next.js. Platform processed 15,000+ grants and $90M+ in 2024. Integrated conversational AI for DAF funding and charity grants โ€” recognised with the 2024 FinTech Breakthrough Award.

Stack: Next.js Conversational AI FinTech

InvoiceMart โ€” RBI-Regulated TReDS Platform

2022

An Axis Bank ร— mjunction joint venture serving 28,000+ MSMEs. Built auction and invoice workflows with bids settled within 24 hours. Production-grade financial compliance and real-time bid matching at scale.

Stack: React.js Node.js RBI-Regulated

GoGiv โ€” Enterprise SaaS Platform

2023

Multi-tenant SaaS with RBAC and white-label support for enterprise clients including Naukri.com. Achieved 99% uptime on AWS.

Stack: React.js Node.js AWS RBAC
Additional works: Wyvate โ€” Food delivery & restaurant management with real-time tracking (99% uptime). Youshd โ€” Influencer monetisation platform integrated with Instagram (25% increase in active users).
Chapter V

Certifications & Credentials

Table 2. Professional certifications by issuing body and date.
  1. IBM โ€” Build RAG Applications: Get Started Mar 2026
  2. Anthropic โ€” AI Fluency: Framework and Foundations Mar 2026
  3. MongoDB โ€” Building RAG Apps Using MongoDB Mar 2026
  4. Anthropic โ€” Introduction to Model Context Protocol Mar 2026
  5. IBM โ€” Develop Generative AI Applications Mar 2026
  6. Anthropic โ€” Model Context Protocol: Advanced Topics Mar 2026
  7. Anthropic โ€” Building with the Claude API Mar 2026
  8. IBM โ€” Data Analysis with Python Feb 2019
Chapter VI

Academic Background

MCA โ€” Master of Computer Applications
Vellore Institute of Technology
CGPA: 8.65 / 10
BCA โ€” Bachelor of Computer Applications
Ranchi University
Chapter VII

Correspondence

The author welcomes inquiries regarding new projects, collaborations, and opportunities. Correspondence may be directed to any of the following.