Full-stack developer building AI-powered products.
I'm Anshul Gautam — I design, build and test complete web applications, from React frontends to Node.js APIs and PostgreSQL data models, with a growing focus on generative AI.
An engineer who ships the whole system — and proves it works.
I'm an MCA student with a foundation in software development, data structures, database management and object-oriented programming. My interests sit at the intersection of full-stack development, artificial intelligence and cloud computing.
Rather than collecting tutorials, I build and test practical software. My flagship project, AI-CRM, is a complete SaaS application — frontend, API, database schema, role-based auth and AI features — validated end-to-end with an automated Playwright suite where all 188 tests pass.
One project, built properly.
Instead of ten half-finished demos, I invested in one production-grade application. Keep scrolling — it's running on the machine below.
AI-CRM — an AI-powered CRM SaaS
What it is
A complete customer-relationship-management SaaS: teams manage customers, leads, a drag-and-drop deal pipeline, tasks, activity timelines and file attachments, with analytics dashboards and AI-generated interaction summaries & follow-up suggestions layered on top.
The problem it solves
Small sales teams lose context between conversations — who said what, which deal is stalling, what to do next. AI-CRM centralizes every interaction, then uses an AI provider layer to turn raw activity history into actionable summaries and next-step suggestions, so follow-ups don't rely on memory.
My contribution
- Developed the application end-to-end: React frontend, Express API, PostgreSQL schema via Prisma, and workflow-based features.
- Implemented JWT authentication with access + httpOnly refresh tokens and role-based access (Admin / Sales Manager / Employee visibility scopes).
- Built a pluggable AI provider service — mock by default, switchable to Anthropic or OpenAI via environment config.
- Owned API & workflow testing: automated end-to-end suites with Playwright — 188/188 tests passing on a single worker.
- Containerized the full stack (db + api + web) with Docker Compose, including auto-migrations and seeding.
Technical implementation notes
- Modular backend — feature modules (auth, customers, leads, deals…) with shared middleware for auth, authorization, Zod validation and error handling.
- Typed data layer — Prisma over PostgreSQL, with schema migrations and seed scripts; monorepo managed with npm workspaces.
- Modern frontend state — TanStack Query for server state, Zustand for auth/theme, React Hook Form + Zod for forms, dnd-kit for the pipeline board, Recharts for analytics.
- Data exports — ExcelJS (.xlsx) and PDFKit (.pdf) report generation.
Tools I work with, grouped by what they're for.
Education & credentials.
- Organizing committee volunteer — Amar Ujala Jyotish Maha Kumbh (2024)
- Volunteer — 38th National Games, Uttarakhand
- Volunteer — GRAFEST (2023, 2024)
Let's build something worth testing.
Open to internships, junior full-stack roles and AI product work. The fastest way to reach me is email.