Sunil Jadhav | AI Knowledge Systems Architect

Business Impact

Strategic outcomes, not vanity metrics

Results delivered through AI-native knowledge systems at CleverTap

18,000+

Questions resolved by AI

Answered by AI-powered knowledge systems across the docs

4,165

Support hours deflected

Customer queries resolved without a ticket

₹1.5 Cr+

Support cost avoidance

Including replacing a third party support bot vendor

68.92%

Fewer SDK tickets

After the developer-first SDK docs redesign

188%: Quarterly throughput growthMedian turnaround cut from 21.5 to 9.1 days
5X: Productivity increaseGPT-driven content lifecycle automation
172%: Docs engagementSDK landing page visits after the revamp
800%: UX writing capacityThrough internal upskilling, not headcount

Strategic Value Proposition

What I Deliver

●Knowledge infrastructure that scales with product complexity

●Support cost reduction through intelligent automation

●Developer activation systems that compress time-to-value

●Content operations that multiply team output without headcount growth

The Future I'm Building

→Knowledge that updates itself when code changes

→Documentation drift detection through AI-powered governance

→Support queries that continuously train knowledge systems

→Developer friction surfaced before it becomes a ticket

This is not content operations. This is knowledge infrastructure as competitive advantage.

Career

Experience

CleverTap

Cloud customer engagement SaaS platform · Mumbai

2019 to present

Progression

Director, Technical Writing
2025 to present
Associate Director, Product Delivery
2022 to 2025
Documentation Manager, Associate Director
2021 to 2022
Team Lead, Senior Manager
2019 to 2021

Recognition

Most Valuable Player in FY2026, CY2023, H1 2024, and 2022

Selected impact

Leads documentation and UX writing across 21 product areas. Resolved 1,286 tracked issues in FY2026 while cutting median turnaround from 21.5 to 9.1 days and lifting quarterly throughput 188%.
Built and product managed the Ask AI support bot end to end, replacing a third party vendor (about ₹26.46 lakh a year saved). Implemented Kapa AI across the docs: 18,332 questions resolved and 4,165 support hours deflected.
Built a suite of Claude skills that draft user guides, developer guides, and API references to house style, with technical claims verified against the source code.
Co-created DriftAudit, which compares code against current docs and flags stale content before users hit it. Scaled from a 7 issue pilot to 78 proactive fixes per quarter.
Built an agentic release notes system (Confluence to ReadMe) and an AI reviewer grounded in the style guide, with a metrics layer tracking draft quality.
Introduced the team's first AI tooling, including Peer Review GPT, and cut new hire ramp up from six months to two with a UX writing training and certification path.
Revamped the SDK landing pages: integration tickets down 68.92% and page visits up to 172%.
Created llm.txt so CleverTap documentation can be crawled and ranked by AI agents and answer engines.

Earlier career

2018 to 2019

Freelance

Technical Writer and Trainer

Trained engineering staff at Mitsubishi Fuso Motors on technical report writing, onsite in Kawasaki, Japan.

2014 to 2018

SmartStream Technologies

Senior Technical Writer

Introduced agile writing and standardized documentation across financial domain products.

2010 to 2014

Amdocs

Senior Technical Writer

Documented enterprise applications and patch bundles. Received a Most Valuable Player award.

Leadership Principles

What I am, and what I am not

The standard I hold myself to as a technical content leader.

What I am

A leader who is also an individual contributor. I set direction for the content function and still write, review, and hold the quality bar myself. Staying close to the work keeps my plans realistic and my feedback credible.
An owner of content as a system. Information architecture, taxonomy, style, governance, and publishing pipelines are designed together, so quality scales with the product instead of depending on heroics.
A partner to Engineering and Product. Content is planned with the roadmap, verified against the code, and shipped with the release. A feature is not done until people can understand and adopt it.
Accountable to business outcomes. I measure content by what it changes: support deflection, developer activation, time to value, and cost per content unit. Those are the numbers I report to leadership.
A builder of people. I hire for judgment over polish, give writers clear growth paths, and develop the next generation of leads. A strong team is the only plan that survives the next reorg.
Ready for readers that are machines. AI assistants and answer engines now read the docs before customers do. I structure content so it stays accurate and findable for both.

What I am not

A bottleneck. I empower my team to take decisions and build system guardrails around them: style linters, review gates, and clear ownership. Good calls happen without waiting for my approval.
Someone who hands AI the pen and walks away. AI drafts and people decide. Every technical claim is verified against its source, and prompting skill is never a substitute for understanding the product.
A ticket taker. Requests are prioritized against OKRs and customer impact, not by arrival order or by who asked loudest. Saying no to low value work is part of the job.
A manager of vanity metrics. Page counts and word counts say nothing about whether content works. If a metric cannot change a decision, I stop reporting it.
Attached to how things have always been done. Legacy content, tools, and processes get audited and replaced when they stop serving readers. The biggest wins usually hide in the content everyone avoids.

Selected Work

Projects & Innovations

Building tools and systems that amplify human capability

CleverTap Documentation Automation

Five Claude Code skills (partner docs, API references, developer guides, and user guides) sharing one workflow: Jira intake, structured drafting, claims verified against the live codebase, and one-click publishing to ReadMe, with a learning loop that improves the skills from real usage.

Claude Code SkillsCodebase VerificationDocumentation Pipeline

Impact: End-to-end test caught 3 defects before rollout; live in production

Internal tooling, live at CleverTap

Resume MCP ServerNew

My resume as a remote MCP server. Add it as a custom connector in Claude and ask about my experience, projects, and skills directly. Answers come only from approved resume data.

MCPDockerDigitalOcean

Impact: Live at mcp.suniljadhav.in

Try it ↗

ShivSamvad (शिवसंवाद)

A full-stack bilingual AI persona chat PWA. FastAPI and MongoDB Atlas backend, React 19 frontend, real Claude integration, auth, and offline install.

FastAPIReact 19ClaudePWA

Impact: Full-stack AI product, built solo

Live ↗

This Website's AI Chatbot

The "Ask About Sunil" widget in the corner shipped through three full rebuilds: a custom Supabase and Groq backend, then a Dify-powered proxy, then Dify's own hosted embed. Each rebuild replaced the last once a simpler path proved better.

Supabase Edge FunctionsDifyIterative Architecture

Impact: Rebuilt twice, shipped simpler each time

Try it, bottom right corner

See all projects →

Core Expertise

Building AI-native knowledge infrastructure at scale

AI Knowledge Infrastructure

AI Workflow Design
Knowledge Systems Architecture
Developer Experience
Content Strategy
Documentation Governance & Code-Drift Control
Support Deflection Systems
Content Lifecycle Automation
Cross-Functional Program Leadership

Product Experience

UX Writing Leadership
Product Copy Integration
Embedded Content Workflows
Product Adoption Systems
Onboarding Ecosystem Design
Guided Content Strategy
Video Learning Architecture

Leadership & Operations

Team Development & Mentorship
Agile Content Operations
Strategic Planning & OKRs
Stakeholder Alignment
Resource Optimization

Technical Domain

API & SDK Documentation
Developer Onboarding Systems
Knowledge Base Architecture