ABHISHEK JAIN: A story in five cities

SCHOOL: Where it all started

I grew up in Aurangabad, the City of Gates, and spent ten years at Nath Valley School. It is where I found the three things I still chase: a stage, a good argument, and something to build.

THEATRE: Et tu? I played Brutus

My first big role was Brutus in Julius Caesar. In that spotlight I learned that a room will follow you if you believe the line yourself, and I have been performing in one form or another ever since.

STREET PLAYS: Earning the crowd

Then came nukkad nataks and Hindi drama in open courtyards: no stage, no lights, no captive audience. If you lost people, they simply walked away. It is still the best lesson in storytelling I have had.

DEBATE: Learning to argue well

Debate taught me to build a case fast, find the weak point in the other side's, and, hardest of all, change my mind out loud when they were right.

NEWSLETTER: Deadlines, the first time

On NVS Echo, our school newsletter, I learned what a deadline feels like: chasing the story, rewriting it until it was tight, and getting every issue out on time.

CODE: Where I learned to code

School is where tech clicked for me. I started with QBasic, moved on to C++ and object-oriented programming, and somewhere between the first program that ran and the hundredth bug, I fell in love with programming and building things.

SPORT: State-level badminton

On court, I played for Aurangabad at the Maharashtra State-level Championships. Badminton is a game of quick reads and quicker recoveries, which turned out to be good training for everything after.

NIT CALICUT: Four monsoons at NIT Calicut

In 2019 I moved almost a thousand kilometres south to NIT Calicut. Four monsoons later I graduated fourth in a class of 40, with a much bigger idea of what I wanted to build.

8.28 CGPA, Rank 4 / 40

MACHINE LEARNING: Falling for machine learning

College gave me the maths; curiosity did the rest. Through college I taught myself machine learning and AI, from linear algebra and statistics to building models of my own. Years later, that is exactly the work I ended up shipping.

ISTE NITC: Technical Head, ISTE

As Technical Head of ISTE I ran everything technical for the club with a six-person team. We built the club's websites, automated the repetitive work behind every event, and set up the pipelines that kept it all running. The biggest test was a nationwide mock JEE for more than 4,000 students. That year, ISTE was voted the best club on campus.

Websites. Automations. Pipelines. Mock JEE for 4,000+ students. Best club, 1st of 14

WAYANAD: A flood-warning plan that got adopted

NITC picked me for a Government of Wayanad project on rainfall-triggered flood warnings. I modelled where rain gauges should go across three flood-prone towns, and the district disaster authorities adopted the plan.

850K+, residents covered

COMPETITIVE PROGRAMMING: Sixty contests a year

As Marketing Head of CPHUB, our competitive programming club, I helped keep around 60 contests a year running, busiest in placement season when everyone needed the practice. I also built our sponsorship pipeline, from cold pitch to final payment.

60+, contests hosted per year

RAGAM & TATHVA: Learning to sell

Ragam and Tathva are two of South India's biggest college fests, and they taught me to sell. We pitched brands like Royal Enfield, SBI, Sony and Realme for fests that drew more than a lakh people a year.

₹24L, raised in sponsorship

1 lakh+ combined footfall

FOUNDER'S OFFICE: My first product lesson

At dRisk (Kandle) I worked with the CEO and CTO and rebuilt the sign-up flow by cutting every step that did not earn its place. Completion rose 40%, weekly sign-ups grew by 400 and retention went up 15%. Less really was more.

40%, lift in sign-up completion

BEYOND CLASS: Evenings by the sea

Outside class I moderated at DebSoc, acted with the street play team and Drama Club, played badminton for NITC at the Inter-NIT Championships and taught science and math to 250+ underprivileged students. Late at night it was Dota 2: ₹65,000 in prize money across 12 tournaments.

FIRST OFFER: An internship, then a ticket north

In the summer of 2022 I interned at Salesforce and automated a clean-up chore that used to take weeks, turning it into a same-day pipeline that saves about $120K a year. It ended with a pre-placement offer, and a move to Hyderabad.

$120K, saved per year

SALESFORCE: Day one, badge in

I joined Salesforce's Developer Productivity and AI Reliability org straight out of college. The job: build the platforms, and soon the AI tools, that more than 3,000 engineers use every day.

3,000+, engineers using it daily

AUTOMATION: Four days to six hours

Every production change waited on a chain of manual sign-offs, often for days. I automated that chain across 350+ requests a day, then folded budget, compliance and legal checks into it, giving client teams back about 120 hours a month.

94%, faster approvals

ORCHESTRATION: Onboarding in weeks, not months

New engineers used to spend months assembling their setup by hand. I turned that ad hoc configuration into reusable templates and swapped environment tickets for on-demand tooling. Onboarding went from four months to two weeks.

95%, faster engineer onboarding

RAG: Where college ML met production

I launched a RAG assistant that answers repeat support questions from a 12,000+ document index. It was one of Salesforce's first centralised AI products for developer teams, proven on 7 use cases and then scaled to 60+ teams.

$500K, saved per year (1,400 hours a quarter)

MCP: Plain English in, infrastructure out

Next I built an MCP-based service that lets AI agents turn plain-English requests into Terraform for 35+ teams. Setting up a new developer went from three or four days to under an hour.

35+, teams self-serving infra

AI RELIABILITY: Teaching AI to explain its mistakes

When an AI answer is wrong, the first question is why. I designed an eval framework that scores retrieval and response separately, so 800+ engineers could tell missing context from model error on every failure.

14%, better answer accuracy

AI PLATFORM: LLM access in under an hour

Getting access to an LLM used to take eight days of tickets. I launched a self-service gateway where teams set their own rules inside IAM guardrails, and provisioning dropped to under an hour.

99.5%, faster provisioning

PATENT PENDING: Catching bad deploys early

I designed a patent-pending validation layer that stops misconfigured Helm deploys before they ever ship. Today it quietly guards more than 30,000 builds a day.

30,000+, builds protected daily

PRODUCT: Listening my way to multi-cloud

Interviews and telemetry kept pointing at one blocker: we only supported AWS. I wrote the specs that took the platform to Google Cloud and Alibaba Cloud, and replaced scattered GitHub requests with a single intake portal.

40%, growth in internal users

80% faster request turnaround

ON STAGE: Back in the spotlight

The kid who played Brutus ended up on much bigger stages. I spoke about developer productivity at Dreamforce and TrailblazerDX, Salesforce's flagship conferences, to more than 2,000 people.

2,000+, live attendees

RECOGNITION: Engineering Excellence Award

I won the Engineering Excellence Award after two nominations, was fast-tracked to MTS, and was rated in the top 10% in all three review cycles, with two spot bonuses along the way.

CO-FOUNDER: Curio, and one painful number

On the side, I co-founded Curio to grade handwritten math and science answer sheets against CBSE, ICSE and IB rubrics. It began with 45+ teacher interviews and one painful number: four or more hours a day lost to grading.

45+, teacher interviews

PRODUCT: Marking the working, not just the answer

A vision model splits each sheet into steps before OCR, so we grade the working, not just the final line. Every teacher correction teaches the model, and teachers went from accepting 62% of our grades to 78%.

78%, teacher acceptance, up from 62%

TRACTION: From pilots to paying schools

With founder-led sales and pricing tied to the teacher time we saved, we turned six-week pilots into paid annual contracts: six campuses onboarded, four paying schools and a pipeline of 25+ more, all in six months.

₹4L ARR, from 4 paying schools

ISB: Back to school

Now I'm at ISB's Mohali campus for the PGP, specialising in Information Management. After years of building products, I'm here to learn the strategy behind them.

TECH STRATEGY & AI SIG: Teaching GenAI to my classmates

As a founding member of the Tech Strategy & AI SIG, I built and ran our applied GenAI teaching series, and hosted CXOs on how companies are actually turning AI into business value.

150+, members in the series

CONSULTING: Pricing a cloud-cost startup

For my ELP project I'm redesigning pricing and building a 180-day go-to-market plan for a B2B cloud-cost startup, benchmarking six competitors and interviewing 30+ C-suite leaders along the way. The client stays unnamed under NDA.

~30%, targeted revenue uplift

BUILDING NOW: What's next

Whatever I build next will sit where these threads meet: AI, product and a good story. This card updates as it ships.

FollowUpOS

An AI inbox for small businesses like salons, bakeries and tutors. It reads customer messages, ranks conversations by urgency and how likely they are to turn into business, and drafts replies the owner approves before sending.

Next.js, AI triage, Drafted replies

FrontDesk AI

A voice receptionist for clinics. It answers patient calls using the clinic's own policies, books appointments, collects a short intake, and hands the doctor a brief in Claude.

Voice AI, MCP, Claude

Armoire

A digital wardrobe with an AI stylist. Photograph each piece once and it is tagged by colour, category, fabric and season; then Armoire builds outfits from what you already own and points out what's missing before you shop.

AI stylist, Image tagging, Outfit generation