Hardhik Potharaju
Hardhik Potharaju

Hardhik Potharaju

Senior PM · Adobe

Taking AI from prototype to production on a $6B enterprise platform.

Adobe$6B platform2 Innovation Awards

I'm Hardhik Potharaju, Senior PM at Adobe with 10+ years across enterprise SaaS, FinTech, and B2B platforms. I ship platform products: data models, RBAC, multi-tenant controls, and 2 governance-cleared AI systems in production on a $6B renewals platform with 2 more in pilot, built without dedicated engineering. Across Deloitte, LTI, and Cognizant, I launched regulated fintech, B2B marketplaces, and HIPAA payer systems. 2 Adobe Innovation Awards. SDA Bocconi MBA.

Companies

AdobeDeloitteL&T InfotechCognizant
Read the work → Ask my portfolio →Get in touch →

Selected work · 6 cases

Production work and the design calls behind it.

See all cases →

~800

field fills · 26 min · 99.8% accuracy

№ 01
PRODUCTIONShipped, in use globally

Pipeline Automation

Took a year-old, ~800-field manual revenue-ops task per run across three regions to a reviewed automation cleared by Legal, IT and enterprise security review, 87% time reduction, in production globally.

Python · Playwright · CSVRead →
Copilot Agents for Product and Sales-Ops Teams№ 02
PRODUCTIONAdobe, in use by 60+ colleagues

Copilot Agents

Two copilot agents serving Adobe Product and Sales-ops teams, 50+ PRDs, user stories, and release notes generated through agent loops, both agents cleared Adobe Legal and AI review.

LLM agents · PRD automation · User-story synthesisRead →
Storyboard AI, Script-to-Visual Generator№ 03
BETASubmitted to Adobe Incubator

Storyboard AI

Turns scripts into visual storyboards with character consistency across frames, submitted to Adobe Incubator and beta-tested by 5 marketers for real-world campaign ideation.

OpenAI image gen · Scene analysis · Character consistencyRead →

NL → SQL

56-column schema · SELECT-only enforced

№ 04
PROTOTYPEPrototype, governance-cleared design

NL Query Layer

A natural-language query layer over pipeline data where no records ever reach the model, a 5-layer SQL guard enforces read-only access in code, not in policy.

Python · SQLite · Internal LLM APIRead →

Plan → Query → Execute → Review

Human gate · Cannot auto-commit

№ 05
PROTOTYPEPrototype, agent design cleared internally

Agentic Operations Loop

A planning agent that reasons over pipeline state, proposes write actions, and hands off to a human at every consequential step, human review is not configurable out.

Python · SQLite · Internal LLM APIRead →
PawSearch, Multimodal RAG on Home Camera Feed№ 06
EXPERIMENTHobby project, user pull for production

PawSearch

A full multimodal-RAG pipeline that searches home video by natural language, no tags, no keywords. 5+ users have asked for a production release.

FFmpeg · Gemini Multimodal Embeddings · ChromaDBRead →

The masthead · what I'm hireable for

Skills, with proof attached.

No skill bars. Every claim links to the case where the work was done.

Platform strategy & roadmapOwned the data model, RBAC, and multi-tenant controls for a $6B renewals platform across AMER, EMEA, and JAPAC. CES raised to 4.5/5, ~108 minutes saved per user per week.See ↗Agentic AI in productionShipped copilot agents to production for Adobe Product and Sales-ops teams. 50+ PRDs, user stories, and release notes drafted through agent loops, used by 60+ colleagues, cleared Adobe Legal and AI review. 2 Adobe Innovation Awards.See ↗Applied AI under complianceNatural-language query layer designed for enterprise Legal, IT, and Security review. Deterministic runtime, 5-layer SQL guard, no customer records in the model.See ↗Fintech & regulated productRBI Payment Aggregator strategy for a Singaporean digital wallet's India launch. KYC completions +80%. Founding PM on a B2B lending marketplace that hit Rs 1,500 Cr GMV with KYC/KYB automation.See ↗Enterprise transformation under regulationGDPR data-privacy standardisation across four global regions at Adobe. HIPAA payer-system automation on FACETS and Trizetto for Anthem and Emblem Health. GxP clinical packaging transformation for Catalent across USA, EU, and APAC.See ↗Multimodal AI productsStoryboard AI, submitted to the Adobe Incubator and beta-tested by 5 marketers for live campaign ideation. PawSearch, an end-to-end multimodal RAG pipeline on home video.See ↗

The masthead · career

Where I've shipped, where I learned.

Select a company to see the role, what was owned, and the outcome. Education sits in the same timeline.

Bangalore, India · Feb 2026 → Present

Senior Product Manager, Platform

Adobe

Adobe
2 AI live · 2 in pilot~57 hrs/wk returned2 Innovation Awards

Built and shipped 2 governance-cleared AI systems in production on top of a $6B renewals platform (pipeline automation and copilot agents) and 2 more in pilot (a natural language query layer and an agentic operations loop). No dedicated engineering.

  • Pipeline automation tool (Python + Playwright): ~800 field fills per run in ~26 min at 99.8% accuracy, returning ~57 hrs/wk across AMER, EMEA, and JAPAC. Full audit logging and a human-review gate before any data commit.
  • Cleared Data Security, Legal, and IT governance for production rollout using AI-assisted development.
  • MS Copilot RAG agents (Policy Q&A + PRD automation) cutting response time from 2 hrs to 2 mins; rolled out to 60+ colleagues after Legal and AI governance review.
  • Designed and ran AI enablement (training decks + workshops) for 60+ colleagues to scale adoption.
Agentic AIPlatform PMNL to SQLGovernance0 to 1

Selective engagements

Advisory

I take on a small number of advisory engagements outside my full-time work. Three areas where I have the deepest track record.

01 · Agentic AI deployment

Helping teams take AI from prototype to something that clears governance and reaches real users inside enterprise constraints. Drawn from shipping two governance-cleared AI systems in production on a large enterprise platform, plus two more in pilot, built without dedicated engineering.

02 · Product strategy

Positioning, roadmap, and product-market decisions for FinTech, B2B and platform products. Drawn from a decade across Regulated fintech and Banking, Enterprise, B2B marketplaces, and multi-tenant platform builds.

03 · Digital transformation

Taking legacy enterprise workflows to modern, automated, AI-assisted systems. Drawn from transformation work across large consulting and enterprise environments at Deloitte, LTI, and others.

I work with a small number of teams at a time. If the problem fits, I'm worth a conversation.

Start a conversation →or email directly →

The agent desk · live

Ask my portfolio, instead of skimming it.

A retrieval agent grounded only on the six cases and the seven smaller bets on this site. Ask about a skill, a domain, a trade-off, or paste a JD and see what fits. It cites the case it pulls from and refuses to invent facts outside the corpus.

Try a question

Open the ask interface →