About

Shahil Jha

Full-stack & AI engineer. Noida, India. I build the internal systems companies run on.

Noida · India

Shahil Jha

Full-stack & AI engineer

  • Meta Full-Stack Developer
  • Generative AI with Large Language Models
  • Docker Certified Associate

I build software for the unglamorous parts of a business — the stock count, the settlement file, the renewal nobody remembered, the incentive sheet that took three days and still caused an argument.

I work as a senior developer for a national apparel retail group, where over the past year I have shipped twelve internal systems: sales analytics across more than fifty stores, physical stock audits reconciled against the ERP, marketplace payments tied back to orders, contracts with their renewal chains, hiring, and a unified customer view across online and offline.

The AI work grew out of that, not the other way round. Once you have a warehouse with millions of rows in it, the obvious next thing is to let people ask it questions. The non-obvious part is making sure the answers are correct — which is where most of my effort now goes, and why there is a whole page on this site about what stops an AI from stating a number it did not get from your data.

I take on a small number of outside projects. If your problem is a spreadsheet that has quietly become critical infrastructure, or a report that takes a week, or an ERP that does everything except the one thing you need — that is the work I do.

B.Tech, Information Technology, IIMT College of Engineering, Greater Noida, 2022.

Where I have worked

2026 — present

Senior Software Developer

National apparel retail group (remote)

Twelve internal systems across finance, stock, sales, HR and legal operations. Every case study on this site comes from this work.

2023 — 2025

Full-stack Developer

Venture studio

A large consumer marketplace platform on a distributed services design — Node.js services with Redis acceleration, background workers, scheduled automation, S3-backed media pipelines and Cloudflare hardening. An HR automation suite covering salary workflows, attendance, employee lifecycle and real-time notifications, which reduced manual HR work by 40%. A restaurant ordering and billing platform with event-driven order handling and multi-device synchronisation. A layered application with centralised role-based access control and event logging.

2022 — 2023

Full-stack Developer

Immigration services

A flight booking platform with search, fare comparison, booking and seat selection. An airline integration service pulling real-time pricing and availability from multiple airlines, with caching and query optimisation. A client management system covering visa stages, ticketing, follow-ups and payments.

2022

Backend Developer

Logistics

Backend services for a national parcel and shipment management system — intake, sorting, dispatch and delivery updates.

What I have built

Twelve systems, one verified fact each. This proves the number rather than asserting it.

Stocktake

Store stock screening and physical-audit reconciliation, with the recovery policy applied as written.

Customer Atlas

Unified customer intelligence over 2.35 million profiles, with a grounded AI analyst on top.

ProfitLens

Store-level profit and loss, live, with automatic alerts on loss-making stores.

Settle

Orders reconciled against settlements and lost claims, with a self-hosted AI analyst on top.

Daybook

Self-serve daily sales and returns reporting across marketplaces, rebuilt for speed.

Shortlist

An employee referral portal where every résumé is scored against the job description automatically.

Agreement Vault

Every agreement, its documents and its renewal history in one place, with reminders that arrive early.

Opsdesk

Around ninety screens replacing spreadsheets and email threads across seven external platforms.

Salesgrid

Sales analytics you can hand to fifty store managers without giving away your margins.

Payout

Store incentives calculated the same way every month, for every store.

On Hand

Live stock on hand, with a per-SKU ageing ladder by size and store.

How a project actually runs

  1. 01

    One call, no charge

    Twenty minutes. You describe what is slow, manual or unreliable. I tell you honestly whether software is the answer.

  2. 02

    A written scope, a fixed price, a date

    Agreed before I start. I do not begin on an estimate.

  3. 03

    You see it working in weeks, not quarters

    Something real and usable early, then the rest.

  4. 04

    Handover

    Code, database, documentation, deployment. On your servers, on standard technology. Nothing you rent from me.

What I don’t do

I don’t build mobile apps, brand or marketing websites, or anything where the main challenge is design rather than data. I have also shipped things I would now argue against — one inventory screen got a 3D view that nobody needed and nobody used. I would rather tell you what to cut than bill you for building it.

Certifications

Meta Full-Stack Developer

Coursera

Professional certification in front-end and back-end web development.

Generative AI with Large Language Models

DeepLearning.AI & AWS

The applied engineering behind AI systems, not the marketing.

Docker Certified Associate

Docker

Software that runs the same way on your servers as it does on mine.

MongoDB Certified Developer Associate

MongoDB

Database design and development.

Grokking the System Design Interview

Educative

Designing systems that stay up as load grows.

Every number on this site, and where it came from

₹1.19 crore

Stock variance identified as recoverable in one audit cycle

Stocktake — computed by applying the client’s own written audit policy

99.99%

Barcode match rate across 260,251 reconciliation lines

Stocktake — matched against 334,586 live product codes

2–3 min → 2–5 sec

Report query time, on the same hardware

Daybook — after query and data-model rework

2.35 million

Customer profiles unified across online and in-store

Customer Atlas — 2.7 GB in a single customer-360 table

40%

Reduction in manual HR work

HR automation suite, 2023–25 — documented outcome

153 / 153

Accuracy checks passing on the verified question set

Customer Atlas — six golden-set suites, run before release

Zero

Database queries the AI can write for itself

Customer Atlas — 14 pre-approved data functions; the capability does not exist in the system

12 systems

Built and put into production

Across finance, stock, sales, HR, contracts and customer analytics

1,413,608

Sales transactions analysed, spanning 2021 to 2026

Salesgrid — across 53 stores and 13,081 products

8.5 million

Rows produced by a single join, returned in seconds

Daybook — a 2.4-million-row source fanning out against a 19-million-row table

51 stores

Live profit and loss, profitable versus loss-making on screen

ProfitLens — replaced a Power BI report

52 checks

Automated data-integrity checks on every refresh

Customer Atlas — including assertions that closed months never change value

Tell me what’s slow, manual, or unreliable.

Twenty minutes on a call is usually enough for me to tell you whether software is the answer and roughly what it would cost. If it isn’t, I’ll say so.

Available for freelance builds alongside senior engineering work. One or two projects at a time.