Software & AI Engineer

Anurag
Nigam

> _

I build web applications, AI agents, and AI-powered tools that are fast, elegant, and production-hardened. 4 years shipping at scale.

Let's Talk

Bengaluru, IN

scroll

Anurag Nigam

Software Development Engineer II at SpotDraft, Bengaluru

About

Engineering with craft and conviction

I care about the full picture: not just making things work, but making them fast, maintainable, and a pleasure to use.

Software Development Engineer II at SpotDraft, promoted July 2026 - building complex legal-tech UI systems

Previously founding engineer at Janitri, building real-time labor monitoring for hospitals

Building AI agents and LLM-powered tools for investing, research, and automation

Obsessed with performance: dependency graphs, memoization, topological sorting

B.Tech ECE, Ramaiah University, Bangalore (CGPA 8.29)

Working on billion-dollar ideas on weekends 🚀

GitHub Profile LinkedIn me@anuragnigam.in

Years experience

400+

Hospitals impacted

10+

Features shipped

99%

Latency reduction

Primary stack

AngularReactTypeScriptNext.jsNode.jsAI AgentsLLMsAWS

Experience

Where I've built things

From founding-stage startups to growth-stage product companies. I've owned complex features end-to-end.

SpotDraft

Full-time

Software Development Engineer II

Jul 2026 - Present · Bengaluru, KA

Software Engineer · Aug 2024 - Jun 2026

Perf rating

5 / 5one of few at SpotDraft

Latency

5s → 50ms99% reduction

ARR impacted

$1M+across 50+ workspaces

Led technical discovery, HLD, LLD, and phased delivery of Contract Packets - a multi-document workflow requested by 50+ customers; coordinated six teams and applied agentic development to ship end-to-end in six weeks vs. a six-month conventional estimate
Led technical design and delivery of Signature Block improvements across preparation, DOCX/PDF export, and signing workflows, standardizing rendering and placement behaviour while reducing preparation time by 14%
Designed extensible architecture for Computation Builder including a TypeScript-like DSL, validation, autocomplete, and reusable function framework - enabled customer self-service and generated $100K+ in revenue
Re-architected questionnaire computation using dependency-aware visibility recomputation with dependency graphs, topological ordering, memoization, and targeted invalidation - reducing interaction latency by 99% (5s to 50ms) and cutting large-template completion time from ~30 minutes to 5-10 minutes. The rollout covered 50+ enterprise workspaces representing $1M+ in combined ARR
Improved frontend reliability by resolving workflow correctness, JSON Logic evaluation, and document-rendering issues while strengthening regression coverage
Raised engineering quality across the pod through architecture reviews, reusable library design, mentoring junior engineers, and driving adoption of AI-native practices across planning, development, and testing
Built Angular Agent Skills - a central skills repository that encodes SpotDraft's Angular patterns, architecture conventions, testing expectations, and review standards as reusable context for AI coding agents; works plug-and-play with Cursor, Claude, and ChatGPT. Designed an evaluation framework comparing no-skill vs auto-selected vs forced-skill workflows across output quality, lint/test readiness, and review comments. Shared learnings in engineering channels, which sparked broader team experimentation and influenced adjacent initiatives around feature-flag guardrails and Cursor-in-Jira workflows
Contributed to Neo (SpotDraft's internal AI platform) across both building and documenting - implemented platform behaviors including stop/cancel execution and other orchestration features, created architecture diagrams covering Slack entry points, Mission Control, routing, sandbox execution, and lifecycle states, and participated in design discussions around GitHub auth, RBAC, stale Slack routing, and execution control. The documentation reduced repeated context gaps and helped engineers onboard to Neo more effectively

AngularReactTypeScriptLangium DSLMicro-frontend

Janitri Innovations

Full-time

Founding Product Engineer

Mar 2023 - Apr 2024 · Bengaluru, KA

Hospitals using platform

400+live deployments

p90 performance gain

80%improvement

Bundle size reduction

20%via code splitting

Realate

Side Project

Founder

Aug 2023 - Apr 2024

Hart

Founder

Founder & CTO

Oct 2022 - Jul 2023 · Bengaluru, KA

Accelerator batch

GSF '22Winter cohort

Selected from

Top 200founders nationally

Fountain9

Internship

Frontend Developer Intern

Sept 2022 - Feb 2023 · Remote

Omniflo

Internship

Software Developer Intern

Jun 2022 - Aug 2022 · Remote

Projects

Things I've built

A selection of projects spanning AI automation, blockchain tools, and developer utilities.

Live Product

Zpply.in

AI-powered job application automation

Continuously scans job boards, matches relevant opportunities using rule-based and ML-assisted criteria, and auto-submits applications for scalable workflow automation.

Next.jsDjangoAWSReact QueryTypeScriptClaude AI

Live Demo

Open Source

Voyager

Blockchain transaction explorer

A React-based replica of the Voyager transaction explorer interface with custom infinite scroll, conditional navigation logic, and tooltip-rich transaction detail views.

ReactTailwind CSSInfinite ScrollREST APIs

Source

Open Source

Folder File Browser

VS Code-like file navigation

A TypeScript React app that replicates the file/folder browsing experience of VS Code with an expandable tree structure, configurable file hierarchy, and editor-like UX.

ReactTypeScriptCSS

Source

Full Stack

Chat App

Real-time messaging platform

Full-stack real-time chat application with WebSocket-powered instant messaging, user authentication, and a clean, responsive UI.

ReactNode.jsWebSocketsMongoDB

Source

View all repositories on GitHub

Skills

Technology I work with

From UI systems and state management to cloud infrastructure and observability.

ReactTypeScriptNext.jsReduxReact QueryTailwind CSSNode.jsExpress.jsAngularRxJSFramer MotionWebpackReactTypeScriptNext.jsReduxReact QueryTailwind CSSNode.jsExpress.jsAngularRxJSFramer MotionWebpack

ClaudeOpenAIAI AgentsLLM integrationsCursorPrompt EngineeringAWSFigmaPostgreSQLMongoDBJestDatadogClaudeOpenAIAI AgentsLLM integrationsCursorPrompt EngineeringAWSFigmaPostgreSQLMongoDBJestDatadog

Frontend

ReactNext.jsTypeScriptTailwind CSSFramer MotionReduxReact QueryAngularRxJS

Backend & Infra

Node.jsExpress.jsAWS (EC2, S3, Lambda)PostgreSQLMongoDBFirebaseREST APIs

AI & LLM

ClaudeClaude CodeChatGPTOpenAICodexAI AgentsAgent SkillsPrompt EngineeringRAGMCPAgentic Workflows

Tools & Process

GitFigmaJestDatadogGoogle AnalyticsJIRAWebpackVite

Fundamentals

DSAOOPSystem DesignPerformance OptimizationResponsive DesignAccessibility

Writing

Thoughts on building

Writing about engineering, markets, and things I am building.

All posts

PerformanceFrontendOriginal

From 5 Seconds to 50 Milliseconds: How a Dependency Graph Fixed a Slow Frontend

A contract form with hundreds of conditional fields froze for seconds on every keystroke. It looked like API latency, but the data had already loaded and the main thread was blocked. Here is how dependency-aware recomputation fixed it, and what it taught us about cloning in hot paths.

August 17, 20268 min read

AIInvestingOriginal

An Articulate Coin Flip, or Something Better

I asked my AI if I should buy gold. It gave me entry, stop, and target. Then I asked what it just did. That question led to a method, a logged experiment, and a six-month public scorecard.

August 10, 20265 min read

AIInvestingOriginal

Building My Investor Operating System

I have always been curious about investing, private equity, and venture capital, so I decided to turn that curiosity into a small experiment. I built an AI research desk, ran its first analysis, and chose ICICI Bank as the first live test. More to come.

August 2, 20266 min read

ReactPerformanceMedium

React: Diffing Algorithm, Keys and Why You Should Not Use Index as Key

A deep dive into how React's reconciliation engine works under the hood: why keys matter, how the diffing algorithm uses them, and the three concrete problems that arise when you use array indices instead of stable identifiers.

May 16, 20245 min read

TutorialProductivityMedium