KM

Keshav Malani — AI Product Manager · TPM · Founder

Restlessly Curious

Product leader building enterprise AI, GenAI, and agentic systems — 19+ years across product strategy, roadmap, and delivery, and two companies of my own. At the intersection of enterprise scale and raw building. Always learning, always shipping.

Currently looking: Product Manager or Technical Program Manager roles in AI, and AI-transformation engagements.

Keshav Malani Professional Portrait
Hear Name

The Journey

Started in India, studied Political Economy at UC Berkeley, cut my teeth in strategy consulting at Deloitte, then went to London Business School for my MBA — where I caught the entrepreneurship bug for good.

Since then it's been a parallel life: building ventures from zero (Powr of You grew to 20 countries) while leading complex tech programs at Meta, Koch, and PayPal. I'm the kind of person who runs a GenAI platform for a Fortune 50 by day and debugs a serverless app for a side project by night.

Today I run product for AI systems — enterprise GenAI and agentic platforms, and the governed data foundations underneath them. I'm focused on product management and technical program leadership for AI, and on AI transformation work: helping large organizations actually get value out of it.

Father of two. Board member with the South Asian Berkeley Alumni, bringing Cal alumni together. Vegetarian, board gamer, lifelong reader — always learning new things and meeting new people.

What I Do

smart_toy

AI Product Strategy

Owning vision, roadmap, and delivery for enterprise AI platforms — building AI that ships

hub

Technical Program Management

Orchestrating complex, cross-functional tech programs from concept to completion

groups

Stakeholder & Ops Leadership

Aligning senior leaders, designing org structures and resourcing models, driving operational excellence

analytics

Advanced Analytics & Data

ETL pipelines, BI dashboards, and cloud data warehouses — turning data into decisions

Quick Facts

Education: LBS MBA, UC Berkeley BA
Location: San Francisco Bay Area
Experience: 19+ years
Ventures: 2 operating companies

The Parallel Tracks

Two careers running in parallel for nineteen years. They keep colliding — that's the point.

domain

Corporate Leadership

The Scale Thread

Meta

Meta

Technical Program Manager, contract (Oct 2025 – Present)

Own the roadmap for a conversational AI analytics platform, and built the webapp that retired 200+ nested spreadsheets running the org's real estate project portfolio — data model, ETL pipelines, agent skills and automations end to end, now used org-wide. The analytics platform replaces a static dashboard landscape with on-demand, question-based access to operational data; underneath it sits the governed semantic layer, metric definitions and evaluation framework I own, which scores answers on accuracy and trustworthiness before speed.

more...
Koch Industries

Koch Industries

AI Product Management · AI / ML Technical Program Management (2023 – Oct 2025)

Owned an internal enterprise GenAI platform end to end — orchestration, tooling and connectors that let teams across the business build and ship AI applications, plus the reporting layer that showed leadership where the value and the overlap actually were. Drove vendor roadmaps toward GenAI capabilities, custom connectors and agent protocols, and consolidated fragmented internal tooling onto one platform. Led data, analytics and reporting programs — near-real-time finance data flows replacing month-lagged manual reporting, and US/India delivery teams executing advanced analytics, ML and operations research across supply chain, manufacturing, hospitality and finance. Designed the org structure and resourcing model for a 300+ person analytics, ML and engineering organization — aligning senior leaders on operating processes, managing allocation across the portfolio, and building the capability map that drove placement and hiring decisions.

more...
Meta

Meta

TPM, Legal Tech & Data · Strategic Operations PM (2021 – 2023)

Designed the automated system that triages regulator and law-enforcement data requests — OCR, ML classification, eligibility rules and confidence-tiered escalation against defined SLAs — plus the canonical compliance datasets underneath it, replacing fragmented manual extracts and saving thousands of operational hours. As Strategic Ops PM, secured executive sponsorship for three new governance frameworks: risk prioritization, escalation and leadership review.

more...
PayPal

PayPal

Program Manager & Consultant (2019)

One-year engagement: designed the portfolio governance and program operating model — intake, prioritization, tracking and reporting, with clear ownership and escalation paths — adopted across a technology organization, built through workshops with senior leadership and VPs across business units.

more...
LinkedIn

LinkedIn & Telefónica

MBA-Era Internships (2013)

LinkedIn: MBA intern on the internal consulting team — EMEA org model for FY2014, market-entry strategy, and QBR support. Telefónica Digital: MBA marketing intern on Firefox OS content strategy, focused on South America.

more...
Deloitte

Deloitte

Strategy & Ops Consultant (2007–2012, 5 years)

Clients: Apple, Disney, Oracle, Visa, MPEG LA, Dolby, Agilent. Led IP inspections across 9 countries averaging $19M+ in savings. Built vendor risk assessment for 4,500+ vendors across 21 metrics. Grew internal SaaS from 5 to 19 clients analyzing $109B spend. Recovered $7M from an MFN vendor through pricing compliance analysis.

more...
rocket_launch

Entrepreneurial Ventures

The Builder Thread

NameDrop

Founder (2018 – Present)

Name-pronunciation platform — 200K+ users, 40+ client organizations across government, corporate and education, and 1M+ monthly views, built and scaled with a team of four. Serverless AWS architecture, SSO and identity integration, and web, mobile, announcer, event-badge and LMS surfaces all reading one governed profile and pronunciation layer. namedrop.io

more...

Qurb

Co-Founder (2018 – Present)

A home for small privacy-first tools. UntrackMe strips tracking parameters out of URLs before you share them — built because the alternative is mailing your friends a marketing dossier every time you send a link. qurb.xyz

Powr of You / PY Insights

Co-founder, CEO (2014 – 2021)

Ran a consumer-data and market-research startup for seven years — 20 countries with a team of 9 across the US, UK and India, on GDPR-compliant consent and data-governance infrastructure spanning the EU and UK. Clients: Dynata, NielsenIQ, Behaviorally, Kantar, L'Oréal, Formula 1. Recognition: AI 1000 UK (2018), BIMA 100 (2017), KPMG Shift 100 MadTech (2017). py-insights.com

more...

Enterprise
Rigor

Startup
Agility

join_inner

The Intersections

I believe expertise is most valuable when it's cross-pollinated. My experience in big tech provides the systems thinking needed for startups, while my entrepreneurial ventures keep my leadership approach lean, fast, and experimental.

In practice, this means applying global infrastructure scale to early-stage MVP growth. Using startup resourcefulness to cut through enterprise complexity. Blending technical TPM precision with the empathy and urgency of a founder. Building teams that think like operators and builders at the same time.

How I work

Three calls I've made on AI products, and why I made them.

Accuracy before speed

On the conversational analytics platform at Meta, the obvious metric was response time. I set the evaluation framework to score answer accuracy and trustworthiness first, on top of a governed semantic layer with defined metrics and quality benchmarks — because a fast wrong answer about operational data costs more than a slow right one. Adoption follows trust, not latency.

Consolidate before you build

At Koch, the GenAI platform's most valuable feature wasn't a model. It was the reporting layer that showed leadership where teams were building the same thing twice. That view made the case for consolidating fragmented internal tooling onto one platform. The best roadmap item is often the one you retire.

Ship the data layer, then the AI

Every AI product I've run sat on a foundation somebody had to build first — canonical compliance datasets at Meta, near-real-time finance flows at Koch, one governed profile layer that every NameDrop surface reads from. By the time the AI feature lands, “can we trust this?” has already been answered.

The Stream

Things I'm reading, watching, and thinking about. Cherry-picked from around the web.

Loading stream...

Let's explore an intersection together.

Whether it's AI product and platform work, scaling something from zero, or just a good book recommendation — I'm always open to a human conversation.

share