Stop 01 · 1994 · Childhood

Delhi. Rowdy.

Where the family immigrated to. Born loud, raised loud. Negotiation as a survival skill before I knew what product was.

28.6° N · 77.2° E · origin

Stop 02 · Roots · Every Summer

Punjab. Roots.

Family from a pind near Ludhiana. They moved to Delhi; the roots never did. Wheat, dhol, daal, and a table where everybody talks at once.

30.9° N · 75.9° E · roots

Stop 03 · 2019 to 2024 · Founder

London. Built.

Two companies in five years. UK and the Middle East. CloudApproach and Approachables. Both sold. $575K combined.

51.5° N · 0.1° W · founder

Stop 04 · Now · Rochester

Rochester. Shipping.

AI PM at ESC Partners. Built Maya AI end-to-end. Customer support for utility companies. In pilot with 40+ agents.

43.2° N · 77.6° W · now

Stop 05 · Next · Unwritten

Where next?

Four pins down. The fifth one is not placed yet. If you are building something real in AI, it might land on your map.

?.?° N · ?.?° E · tbd

say hi or ask my AI first

₹2 postage

DELHI G.P.O. 15 · APR · 1995 ★ DEPART ★

born loud, raised loud,learned to negotiate beforei learned to read.

whoever's reading. somewhere on the internet.

₹5 postage

LUDHIANA EVERY · SUMMER ★ ROOTS ★

family loud at the table.bhangra in the bones,daal on every plate.

anyone with a kitchen. come hungry.

£1 postage

LONDON · DUBAI 2019 → 2024 ★ FOUNDER ★

two companies in five years.uk and the middle east.both sold. learned product the hard way.

next operator. ship before perfect.

$1 postage

ROCHESTER · NY NOW · 2026 ★ NOW ★

maya ai. end-to-end.voice · chat · sms · email.70 to 80% first-touch.

the hiring manager. open to ai pm roles.

? postage

YOUR CITY SOMEDAY · SOON ★ NEXT ★

four stops down.the fifth pin is yours to place.tell me where.

the next chapter. shivamsharma2023@gmail.com

AI · PMBUILDERMAYA AIESC PARTNERSVOLLIECLOUDAPPROACHAPPROACHABLESEND-TO-END AI · PMBUILDERMAYA AIESC PARTNERSVOLLIECLOUDAPPROACHAPPROACHABLESEND-TO-END

DELHIPUNJABLONDONDUBAIBUFFALOROCHESTERLUDHIANAUK · MENA DELHIPUNJABLONDONDUBAIBUFFALOROCHESTERLUDHIANAUK · MENA

SHIPPINGFASTAPIREACTBEDROCKDYNAMODBORACLE CCSVOICECHATSMS SHIPPINGFASTAPIREACTBEDROCKDYNAMODBORACLE CCSVOICECHATSMS

UTILITIESSUPPORTAGENT CONSOLEFIRST-TOUCHREAL INDUSTRIESEND-TO-END UTILITIESSUPPORTAGENT CONSOLEFIRST-TOUCHREAL INDUSTRIESEND-TO-END

TWO EXITS$575K COMBINED40+ AGENTS IN PILOTMAYA IN 2.5 MOSIMON MS AI TWO EXITS$575K COMBINED40+ AGENTS IN PILOTMAYA IN 2.5 MOSIMON MS AI

Delhi → Punjab → London → Rochester

··:·· local

shivam sharma.

I build the AI that handles support before a human has to. Maya AI at ESC Partners. scroll for the receipts

Maya AI handles support for utility companies before a human picks up. I built the whole thing — voice, chat, SMS, email, mobile. Before ESC, I started two companies in London and Dubai. Sold both. Then a master's in AI to learn why the models actually work.

shivamsharma2023@gmail.com

Section 02 · Ask about me

Don't scroll.
Just ask.

Trained on everything above. Answers the way I'd answer in person. Under 150 words.

Section 03 · Mind / mouth

Four things I'll fight
you about.

01.

Most AI demos are dishonest. Happy path. Cherry-picked metrics. I won't put a number on a deck I can't defend face-to-face with the engineer who built the thing.

02.

If a metric needs an asterisk, the asterisk goes in the headline, not the footnote. Hiding caveats is how teams lose trust — fast.

03.

Ship ugly on Tuesday. Polish after real users touch it. Six weeks of perfecting something nobody wanted is six weeks you don't get back.

04.

The best place to build AI is the most boring industry you can find. Utilities. Healthcare. Gov. Real users, real edge cases, and nobody there doing a demo day.

Section 04 · Selected work

W O R K 05 / 05 Selected

Things I built
end-to-end.

Five places I owned the product. Maya AI is the headline. Two companies before that — both sold — where I learned what ownership actually costs. Hiring. Payroll. Pricing. The parts nobody puts in a case study.

01 / 05

Maya AI

AI Product Manager · ESC Partners · Jan 2026 to now

AI that handles support for utility companies before a human picks up. Five channels — voice, chat, SMS, email, mobile. Internal agent console plugged into Oracle CCS for the human team.

FastAPIReactAWS BedrockDynamoDBOracle CCSn8n

70 to 80% First-touch coverage

40+ Agents in pilot

Read case study

02 / 05

Assisted Living Locators

AI Product Consultant (Builder) · Remote CA · Oct 2025 to Jan 2026

Matching families with the right care options. Owned the scoring flow, the agent console, and the data layer behind both.

MarketplaceHealthcareB2CAgent tools

2 Surfaces owned

B2C Marketplace

Read case study

03 / 05

Vollie Inc.

Founding Product Manager · Buffalo NY · May to Dec 2025

AI-powered employee coaching, from zero. Built a voice AI that runs the session, moderates the conversation, and writes the report after — sentiment, engagement, commitments.

Founding PM0 to 1Voice AIAI coaching

0 → 1 Founding PM

Voice AI moderator

Read case study

04 / 05

Ain Center

Product Lead · Jan to Dec 2025

Two products at Simon's entrepreneurship center. VentureForge — AI-graded strategy simulations. Validation tool — idea scoring, persona generation, simulated user interviews.

VentureForgeAI simulationValidation toolSelf-serve

Strategy Simulation

Validation Tool

Read case study

05 / 05

CloudApproach + Approachables

Founder · UK + MENA · 2019 to 2024

Two companies. Five years. London and Dubai. CloudApproach: AI for real estate marketing. Approachables: product-led marketing agency. Hired the teams, closed the deals, sold both. ~$5M in revenue influenced across the run.

FounderUKMENASalesExits

Two Exits

$575K Combined

Read case study

Section 05 · How I got here

Three cities. Two exits.
One thread.

LOGBOOK · 6 ENTRIES

2026 · Jan → now NOW

AI Product Manager, ESC Partners

Rochester, NY · Maya AI

Built Maya AI. 80% of first-touch volume handled before a human picks up. 40+ agents on the pilot.
2025 · Oct → Jan 2026

AI Product Consultant (Builder), Assisted Living Locators

Remote, CA

Owned the matching flow and the agent-facing console.
2025 · May → Dec

Founding PM, Vollie Inc.

Buffalo, NY (not Pleasanton)

Wrote the first eval suite. Scoped the first three features. Cut what did not fit.
2025 · Jan → Dec

Product Lead, Ain Center for Entrepreneurship

Simon Business School

Built what students use to apply for programs and grants. Did it while finishing the degree.
2025 · Dec

M.S. AI in Business (STEM)

Simon Business School, Rochester

STEM master's. AI applied to business. December 2025.
2019 → 2024

Founder · CloudApproach & Approachables

UK · MENA

Two companies. Two acquisitions. $575K combined. ~$5M in deals closed across the run.

Section 06 · Reach me

Let's ship
something real.

shivamsharma2023@gmail.com

Coffee in Rochester, or a call from anywhere.

AI Product Manager · ESC Partners · Jan 2026 to now

Maya AI

FastAPIReactAWS BedrockDynamoDBOracle CCSn8n

70 to 80% First-touch coverage

40+ Agents in pilot

Overview

AI that handles support for utility companies before a human picks up. Five channels — voice, chat, SMS, email, mobile. Internal agent console plugged into Oracle CCS for the human team.

Detail

Maya handles 70–80% of inbound support at first touch on one US utility. That’s 100+ emails per manager per day that never reach a human.

What it is

Two surfaces. A customer-facing assistant that handles inbound across voice, chat, SMS, email, and mobile — 26 languages. An internal console for human agents, plugged into Oracle CCS, that loads the context before they pick up.

Most support queries are repetitive. Bill questions. Outage status. Account changes. Maya handles the first touch. When it can’t, it hands off to a human with full context already loaded.

What I built

I was the only AI person at ESC Partners when this started. No handoff from a previous team. No prior codebase to extend.

Built it in 2.5 months. FastAPI backend, React frontend, AWS Bedrock for the LLM layer, DynamoDB for session state, n8n for orchestration, deployed on Render. Oracle CCS was the hardest part — that documentation is not designed for this kind of use.

The moment that made it real: watching an agent use the console for the first time with a live customer. They didn’t have to dig through three tabs to find the account. It was just there.

The number I can defend

70–80% first-touch coverage on one US utility. I can explain exactly how we measure it, what counts as handled, and where the edge cases live.

I will not give you a bigger number than that.

Stack

FastAPI, React, AWS Bedrock (Claude), DynamoDB, Oracle CCS, n8n, Render. Voice via third-party provider. SMS and email through dedicated channels. Mobile via customer widget.

Next case study Assisted Living Locators

Back to the page shivamsharma2023@gmail.com

Delhi. Rowdy.

Punjab. Roots.

London. Built.

Rochester. Shipping.

Where next?

shivam sharma.

Don't scroll.Just ask.

Four things I'll fightyou about.

W O R K 05 / 05 Selected

Things I builtend-to-end.

Maya AI

Assisted Living Locators

Vollie Inc.

Ain Center

CloudApproach + Approachables

Three cities. Two exits.One thread.

AI Product Manager, ESC Partners

AI Product Consultant (Builder), Assisted Living Locators

Founding PM, Vollie Inc.

Product Lead, Ain Center for Entrepreneurship

M.S. AI in Business (STEM)

Founder · CloudApproach & Approachables

Let's shipsomething real.

Maya AI

Overview

Detail

What it is

What I built

The number I can defend

Stack

Assisted Living Locators

Overview

Detail

Vollie Inc.

Overview

Detail

Ain Center

Overview

Detail

CloudApproach + Approachables

Overview

Detail

Don't scroll.
Just ask.

Four things I'll fight
you about.

Things I built
end-to-end.

Three cities. Two exits.
One thread.

Let's ship
something real.