Shaashwat Sharma — Full Stack + AI Engineer

Shaashwat Sharma · Full Stack + AI Engineer

Keep watching, or point at it

Full-stack and AI engineer. I build AI automations, the backends underneath them and the interfaces on top. When I haven't done something before, I find out how, learn it and ship it.

See how I work

50+
Workflows automated: 24H
PRD to shipped product: 2+
Years building

Currently SDE-1 at Core Value Technologies · Noida, India — IST

Keep scrolling. Your scroll is the timeline.

PRD in. Product out.

Plan

Backend

Frontend

Review

Deploy

Hours from requirement to shipped product, backend and frontend. Automated in n8n.

Your turn. The pipeline waits for you.

A human decides.

Step 012 · waiting for a human

Release the build to production?

Decision log2 approved · 0 rework

005Plan reviewedApproved
009Tests passedAuto
011Staging build checkedApproved

Every step and decision on record.

Keep scrolling to hand out the work

Agents on call.

I don’t do every step by hand. I hand jobs to AI agents running in the cloud, each one does its part, and the results come back into my own system for me to check.

01 · A task comes in
I describe the work once, in my own system.
02 · Agents take a job each
AI agents in the cloud pick it up and work in parallel.
03 · Results come back
Each one reports into my system. I check it, then it ships.

Illustration. The six jobs are examples.

Three things, end to end

Give me the problem.

Chapter 01

AI + automation

A requirement goes in. A product comes out.

n8n pipelines that take a PRD through backend and frontend to a finished product
A human approves at every gate that matters, and every step and decision is recorded
Cloud agents and AI harnesses wired into my own systems to get work done

Chapter 02

Systems

Built to hold.

Backends where the business rules live in configuration, not code
APIs, data models and pipelines on Java, Spring Boot and PostgreSQL
Production debugging across Kubernetes and GCP

Chapter 03

Interfaces

Made to be used.

Web front ends in React and Next.js
Motion that explains what the product is doing
iOS apps, designed and built end to end

Selected work

Shipped.

At work · an enterprise insurance platform

Open a row. The first one has a live demo.

Most of this is client work under NDA, so there are no repos or live links to click. What I can do is walk you through the architecture, the trade-offs and what broke — in as much depth as you want.

A rating engine for specialty insurance where pricing logic is authored rather than coded. Underwriters define the calculation steps and the formulas that drive them; the backend validates that definition, executes it against a submission, and returns a result you can trace step by step back to the rule that produced it.
JavaSpring BootRules EnginePostgreSQL
Base rate1200Risk factor1.4Discount %10
1 · base × factor1,680
2 · − discount−168
3 · premium1,512
Illustrative demo. Not client logic or data.
A quoting system where the quote's own fields are user-defined — the shape of the data is configured at runtime rather than fixed in the schema. The interesting problems live in validation, storage and querying: how do you keep a configurable document queryable, and how do you evolve a definition without breaking the quotes already written against the old one?
BackendSchema DesignPostgreSQLValidation
Document generation for enterprise insurance endorsements. Templates and per-field editability are owned by the business as configuration, not code, so a form change is a config promotion rather than a release. Templates render through a logic-less engine to HTML and then to PDF, matching legally-approved layouts exactly.
JavaGotenbergHandlebarsPostgreSQL
An automation layer stitching the delivery flow together from specification through document generation to the finished product, so the pipeline effectively runs itself.
n8nAutomationWorkflowBackend
A microservices order-management system handling 500+ daily transactions, with Server-Sent Events for live order tracking and Docker orchestration keeping uptime above 99%.
Spring BootReactSSEDocker
An interactive React app pulling NASA's open APIs and geolocation to surface cosmic events near you, with a calendar-based visual system and caching that cut API calls while lifting engagement ~30%.
ReactNASA APITailwind CSS

On my own time

Each one is playable

Web app

DairyDesk

B2B dairy wholesale ordering.

Milk · crate2
Paneer · kg1
Curd · tub0

3 lines in this order92 units

Sample items. Not the live catalogue.

An ordering app for dairy wholesale: a buyer side with catalogue and checkout, and an admin console for the business running it.

WebB2BOrderingAdmin console

dairydesk.store →

iOS app

Authenticator

A two-factor authenticator that works offline.

demo@work
••• •••
demo@personal
••• •••

Demo codes. No real account or secret.

A code-generator app in the style of Microsoft Authenticator: add an account, get a rotating one-time code for it.

iOS2FAOne-time codesOffline

GitHub repo →

iOS app

Toggle Tracker

Track your days. See them as a heatmap.

79 active days of 112

Sample data. Click a day to log it.

A productivity tracker for iPhone: toggle what you got done each day and read the pattern back as a heatmap.

iOSProductivityHeatmap

GitHub repo →

Showreel

1:00

Sixty seconds on how I work — AI and automation, the backend underneath, and the interfaces on top.

0:00 / 1:00sound off

Type something you think I can’t do, then press Enter

Never done it?

> a new framework?Learned.
> a new domain?Learned.
> no docs?Figured out.

Research it. Learn it. Ship it.

Where I’ve done it

Experience.

SDE-1
Now
Core Value Technologies
2025 — Present
Noida, India · Backend
- Backend developer on an enterprise insurance platform — own the document-generation services for endorsement forms, where business-owned templates render through Mustache/Handlebars to pixel-accurate PDFs via Gotenberg.
- Built the Nexus pipeline: an n8n-driven workflow that automates the delivery flow end to end.
- Engineered a microservices order-management system (Java Spring Boot + React) — 500+ daily transactions, Server-Sent Events for real-time tracking, and Docker orchestration at 99%+ uptime.
- Own configuration promotion across QA, UAT and production, and debug production issues spanning Kubernetes, GCP Pub/Sub and templating-engine behaviour.
Java / Spring Boot / React / PostgreSQL / Gotenberg / n8n / Kubernetes / GCP / Docker
SDE-1 Intern
Core Value Technologies
2025
Noida, India · IoT & ML
- Built an IoT fuel-monitoring & theft-detection system for UP Roadways — sensors, GPS and ML models cut theft incidents 50%+ at 91%+ anomaly accuracy.
- Shipped a real-time telemetry dashboard monitoring 100+ buses live with React + Node APIs.
React / Node.js / IoT / ML / Dashboards

What I reach for

Toolkit.

Languages

Java
TypeScript
JavaScript
C / C++
SQL