Saffa Ahmad
Résumé

I spent three years building AI products.Now I am learning to choose which ones get built.

Ten-plus enterprise GenAI products shipped at TCS for global clients, discovery through release, with a five-person pod behind me. Currently reading Technology and Business Management at Masters’ Union.

Read a winning caseLinkedIn(opens in a new tab)
Mohammad Saffa Ahmad, photographed in 2026
Mohammad Saffa AhmadGurugram, India

Currently at Masters’ Union, Gurugram

Scroll to the work
02 / 09The position

Four numbers that carry the rest of the page.

10+

GenAI products shipped

enterprise, in production

400+

users on one deployment

70% adoption

30%

client process efficiency

peak improvement

3 yrs

at TCS

two promotions

Nagpur, then Pune, now Gurugramcounting visits…

03 / 09How I decide

The moveyou don’t playis the decision.

Anyone can list what they shipped. Four years of state-level chess taught me the opposite habit: you are judged on the lines you calculated and rejected, because those are the ones that would have lost.

It turns out to be the same skill as prioritisation. A roadmap is a list of moves you declined to make, and the reason you declined them is the only interesting part. So this is the work I would actually defend:

  • 7segments analysed
    3survived the gates

    goSTOPS

  • 6+models benchmarked
    1shipped per use case

    TCS

  • 10+prototypes built
    4became funded work

    TCS

The struck number is the one that mattered.

04 / 09Games played

Work I can show in full

The enterprise work sits under client NDAs and is summarised further down. These are mine to show: the reasoning, the numbers and the parts that did not work.

Two of them won outright, each against a full cohort, in a single day.

01 / 04Winner
goSTOPS one-day challenge

TERM 0

modelled year-one revenue
₹3.9 Crmodelled year-one revenue
primary research
n=80primary research
place
1stplace
The problem
goSTOPS empties out between July and September. Every obvious fix is a discount, which trades margin for a problem it does not solve.
What I did
Proved the quarter is an occupancy problem, not a pricing one. Gated seven candidate segments on three binary tests, ran eighty interviews, and built the winner into a costed three-night product.
What came of it
First place. A modelled ₹3.9 crore of incremental year-one revenue at an unchanged room rate.
  • Segmentation
  • Primary research
  • Unit economics
  • Go-to-market
Read the caseStrategy, research, deck, 2026
02 / 04Winner
Wow! Momo one-day challenge

Wow Masala

place
1stplace
new category proposed
1new category proposed
day, start to finish
1day, start to finish
The problem
Wow! Momo set the brief for the Masters’ Union Consumer Behaviour one-day challenge: a single day, start to finish.
What I did
With my team, proposed a new category for the brand: Wow Masala.
What came of it
First place in the Consumer Behaviour one-day challenge.
  • Consumer behaviour
  • Category strategy
Team entry, 2026
03 / 042026
Automation, built and running

NutriNest Query Router

severity routes
3severity routes
workflow steps
13workflow steps
lines of code
0lines of code
The problem
Support queues where a human reads every ticket have one failure mode: the reading is the bottleneck, and the most urgent ticket sits wherever it landed.
What I did
Rebuilt my TCS triage system small enough to inspect. A model classifies and scores severity, a router sends it three ways, and the customer is acknowledged before any of that runs.
What came of it
Thirteen steps, three severity routes, no code. The enterprise original cut manual ticket handling by forty percent.
  • make.com
  • LLM classification
  • Ops automation
Read the caseDesign and build, 2026
04 / 042026
Side build, playable

Crack the Vault

locks
5locks
lives
2lives
weekend
1weekend
The problem
I wanted to know how far prompt-driven building had actually come, rather than take anyone's word for it.
What I did
Designed and shipped a timed heist brain teaser over a weekend with Lovable, tuning difficulty until losing felt fair rather than arbitrary.
What came of it
A finished, playable game. Built with Lovable, which is the point rather than the caveat.
  • Lovable
  • Game design
  • Rapid prototyping
Play it(opens in a new tab)Concept and build, 2026
05 / 09The middlegame

Three years at TCS, 2022 to 2025

Trainee to Systems Engineer inside three years, two promotions, and a five-person pod by the end of it. Client names are withheld, so each product below is described by the problem it solved rather than who paid for it.

Ten-plus GenAI products shipped into production, client process efficiency up to 30%, and a Technical Excellence Award for the delivery.

  • Chemical manufacturing

    RAG diagnostic assistant

    Plant engineers lost hours tracing fault history across scattered documents.

    20% faster diagnostics, 400+ users, 70% adoption

  • Industrial operations

    Predictive maintenance

    Equipment failures surfaced only after they had already stopped the line.

    15% less unplanned downtime

  • IT support

    LLM ticket triage

    Support teams read and routed every incoming ticket by hand.

    40% less manual ticket handling

  • Manufacturing analytics

    Line-level KPI dashboards

    Plant managers had no live view of where failures were concentrating.

    Real-time failure visibility across production lines

  • Model strategy

    LLM selection framework

    Every use case reached for the largest model, whatever it cost to run.

    6+ models benchmarked on accuracy, latency and cost

  • Enterprise search

    Document intelligence pipeline

    Answers were locked inside scanned documents no search index could reach.

    Document intelligence, embeddings and vector retrieval in production

  • Client development

    Discovery to funded delivery

    Promising prototypes stalled before anyone would pay to productionise them.

    4 prototypes converted into funded engagements

  • The rest of the ten

    The other products sit under the same client NDAs and are left out rather than described vaguely. The résumé covers the scope and the stack behind all of them.

    Product résumé

Figures are the ones reported to the client at handover.

06 / 09Move order

How I got here

Engineering, then research, then enterprise delivery, and now the business side of the same problem. The through-line is building things people actually use.

A filed patent at 20, three published papers by 22, and a five-person pod by 24.

Interactive

Spin through the years

Ten detents, one for each thing worth marking. Drag it, flick it, or use the arrow keys. It carries momentum and settles on a year, the way a real one would.

2018202020212022202220232024202520262026
2018Education

Bharati Vidyapeeth’s College of Engineering

B.E. Computer Engineering, honours in data science. Graduated at 8.49 out of 10.

Drag to spin

  1. 2018

    Bharati Vidyapeeth’s College of Engineering

    B.E. Computer Engineering, honours in data science. Graduated at 8.49 out of 10.

    1
  2. 2020

    Smart Electric Switch

    Patent application 202021019022 filed for a contactless timer switch. Selected for the MHRD National Innovation Contest.

    2
  3. 2021

    First papers published

    Two papers on machine-learning weather forecasting in GIS Science Journal.

    3
  4. 2022

    India Meteorological Department

    ARIMA and machine-learning forecasting on IMD data. Prediction accuracy up eleven percent.

    4
  5. 2022

    Tata Consultancy Services

    Joined the Microsoft Business Unit as a trainee. Promoted inside a month.

    5
  6. 2023

    GenAI delivery

    Built enterprise GenAI products end to end on Azure for global clients.

    6
  7. 2024

    Leading the pod

    Took over a five-person team. Sprints, reviews, and the client stakeholders that came with them.

    7
  8. 2025

    Technical Excellence Award

    Recognised for innovation and delivery leadership. Second promotion in three years.

    8
  9. 2026

    Masters’ Union

    PGP in Technology and Business Management. Core member of the ProdMan Club.

    9
  10. 2026

    Wow! Momo one-day challenge

    First place in the Consumer Behaviour one-day challenge, set by Wow! Momo. Our answer was a new category for the brand: Wow Masala.

    10

Published

  • Weather Prediction Using Machine Learning

    GIS Science Journal, volume 8, issue 12

    GSJ/5750

  • Machine Learning for Weather Forecasting in Python

    GIS Science Journal, volume 8, issue 12

    GSJ/5847

  • E-Learning Personalisation with Artificial Neural Networks

    IJRAR Journal, volume 9, issue 2

    IJRAR22B289

Patented

Smart Electric Switch

A contactless, timer-based switch designed to cut the risk of shock at home.

Application 202021019022

Selected for the MHRD National Innovation Contest, 2020

Recognised

  • Winner, Consumer Behaviour one-day challenge

    Wow! Momo at Masters’ Union, 2026

  • Technical Excellence Award

    Tata Consultancy Services, 2025

  • IELTS Academic, band 7.5, CEFR C1

    British Council, 2025

  • TCS Digital Certification

    Tata Consultancy Services, 2024

07 / 09Next move

Ask my portfolio

I build retrieval systems for a living, so this site should be able to answer questions about me directly. That version is being built now. Until it is genuinely good it stays switched off rather than shipped half-working.

Launching soon

What did he actually ship at TCS?

How does he prioritise a backlog?

Why leave engineering for product?

Soon

In the meantime, everything it would tell you is on this page.

08 / 09Off the board

Away from the work

Four years of state-level chess, which is where the habit of looking several moves past the obvious one came from.

Chess

Four years of Maharashtra state-level play. It is where I learned to look a few moves past the obvious one, which turns out to be most of what strategy work actually is.

Poetry

I write, and occasionally win things for it. An inter-college competition once, and a habit that has comfortably outlasted it.

Go and see (opens Instagram in a new tab)

Mentoring

Peer-learning sessions on AI and DevOps for ten-plus engineers at TCS, which cut new-joiner ramp-up by about two weeks. It is the part of the job I would keep if I could only keep one.

Community

Four years volunteering with the Leo Club of Pune Vibrant, and CSR drives at TCS running technology-awareness sessions.