From structured authoring to AI-native docs
Principal Information Architect with 14 years of experience building documentation systems that scale — from DITA migrations to retrieval-augmented generation (RAG) pipelines and Model Context Protocol (MCP) servers.
Currently leading documentation engineering at Actian: a production RAG chatbot with RAGAS evaluation, an MCP server integrating Claude with the documentation corpus and analytics, a GA4→BigQuery content-gap pipeline, and Jenkins CI automation.
Delivered measurable results: 45% improvement in content discoverability, 50% increase in organic traffic, 70% reduction in engineering onboarding time, and 40% increase in API self-service success.
Beyond writing for people, I make content AI-ready — structuring, chunking, and enriching documentation so large language models can retrieve and answer from it accurately. Designs content architectures optimized for both human readers and LLM retrieval, serving both from a single trusted source.
Things I've shipped
knowflow
MCP server · TypeScriptAn MCP server that makes documentation natively queryable by AI — connects the docs corpus, GA4 content gaps, Jenkins CI, and a RAGAS evaluation layer to Claude in a single conversation.
View repo →Documentation-AI-Assistant
RAG chatbot · PythonA retrieval-augmented documentation assistant that grounds LLM answers in the docs corpus using embeddings and semantic search — turning static pages into a conversational, self-service knowledge source.
View repo →Making content AI-ready
I author documentation for two audiences at once — the developer reading the page, and the model retrieving, reasoning over, and answering from it.
As search shifts from keywords to conversation, "good docs" now means content that an LLM can ground in without hallucinating. I architect every layer — structure, metadata, chunking, and evaluation — so the same source serves human readers and AI assistants with equal accuracy.
Retrieval-First Structure
H2/H3-anchored, self-contained sections sized into ~300-token windows so each chunk stands alone as a clean, high-signal answer for a RAG pipeline.
Semantic Metadata
Product, version, and topic-type metadata plus schema markup that let models filter, disambiguate, and cite the right content with precision.
Machine-Accessible Surfaces
Exposing docs through MCP servers and APIs so assistants like Claude can query the corpus directly — turning static pages into a live knowledge source.
Single-Source, Structured Authoring
Topic-based, docs-as-code content with consistent terminology and controlled vocabulary — clean, predictable input that embeddings and LLMs love.
Grounding & Evaluation
RAG pipelines evaluated with RAGAS — measuring answer relevance, faithfulness, and context recall against a repeatable baseline. Embedding models benchmarked across the corpus. Content-gap analytics surface zero-result queries before a support ticket does.
Continuous AI Feedback Loop
GA4→BigQuery signals and AI query logs feed prioritization, so the content set keeps getting more answerable over time — not just more voluminous.
What I bring to the table
AI & Retrieval Systems
Designing RAG pipelines, MCP servers, chunking & embedding architectures, and semantic search that ground LLMs in trusted documentation.
Information Architecture
Scalable IA frameworks, taxonomy systems, and metadata schemas that improve content discoverability across complex documentation ecosystems.
Content Strategy & Modeling
Comprehensive content strategies, structured authoring models, and reusable frameworks aligned with business objectives and user needs.
DocOps & Automation
Docs-as-code workflows, Jenkins CI/CD, GitHub Actions, and automated validation gates that cut build time and eliminate release errors.
Analytics & Discoverability
GA4→BigQuery pipelines surfacing zero-result searches and high-exit pages to drive data-backed content prioritization.
Team Leadership
Managing cross-functional documentation teams, mentoring writers, and partnering with engineering and product on developer experience.
Professional journey
Principal Information Architect
Actian
Bengaluru, India
Leading the transformation of Actian's documentation ecosystem for Data Platform and Data Intelligence products — pioneering AI-native documentation through RAG, MCP, and analytics-driven content engineering.
Production RAG Chatbot
Built and launched a RAG documentation chatbot combining document embeddings, vector indexing, semantic search, and LLM response generation grounded in the Actian corpus — significantly reducing support dependency at scale.
knowflow MCP Server
Built an MCP server exposing the docs corpus, GA4 data, and Jenkins CI status to Claude, enabling natural-language queries across all three sources in one conversation — with a RAGAS evaluation layer measuring pipeline quality.
Chunking & Embedding Architecture
Designed H2/H3-anchored 300-token windows enriched with product, version, and topic-type metadata for precise filtered retrieval. Benchmarked embedding models using MTEB metrics to optimize retrieval quality.
GA4 → BigQuery Pipeline
Built an analytics pipeline surfacing zero-result searches and high-exit pages as a weekly content-gap prioritization dashboard.
RAGAS Evaluation Layer
Introduced LLM evaluation measuring answer relevance, faithfulness, and context recall to establish a repeatable quality baseline for the RAG pipeline.
CI/CD Automation
Implemented Jenkins-based publishing with automated validation gates, reducing build & release time by 40% and eliminating manual errors.
Portal Rebuild
Led the end-to-end rebuild of the Actian docs portal using Markdoc and a docs-as-code workflow — a reusable architecture spanning multiple enterprise product lines.
Standardized IA
Designed a standardized IA and navigation model improving content discoverability by 45% and reducing duplicate content by 35%.
SEO Program
Led a full SEO program (metadata, URL structure, schema markup, internal linking) achieving 50% organic traffic growth in six months.
End-to-End API Docs
Owned REST and GraphQL API documentation end to end — reference guides, authentication flows, and developer onboarding content — improving API self-service success by 40%.
Team Leadership
Led and mentored a team of 12 documentation professionals across multiple time zones, improving delivery velocity by 25% through authoring standards, review processes, and content templates.
Information Architect
Adobe
Bengaluru, India
Managed content strategy and information architecture for Adobe HelpX, supporting more than one million users per quarter across developer and enterprise documentation.
M&A Frameworks
Built standardized M&A documentation frameworks adopted across multiple acquisitions, reducing engineering onboarding time by 70% and improving consistency across newly integrated products.
Interactive API Docs
Implemented Swagger/OpenAPI documentation with an interactive testing environment, consistently ranked among the top three features requested by developers.
AI-Assisted Migration
Led AI-assisted content migration, taxonomy design, and structured authoring to improve content reuse and reduce long-term maintenance costs.
Analytics Dashboard
Built a content performance dashboard in Adobe Analytics with 99% data accuracy, supporting quarterly content planning decisions.
Personalized Content
Delivered a personalized content plan for checkout and onboarding pages, increasing product adoption by 25% through targeted guides and in-context help.
Senior Technical Writer / Information Architect
ABB
Bengaluru, India
Planned and executed documentation for a full developer ecosystem, including API guides, CLI references, and mobile SDK documentation for Swift and Kotlin.
Research-Driven IA
Restructured the documentation library using card sorting and tree testing, improving navigation satisfaction scores by 40% in user surveys.
DITA Migration
Led the migration of legacy Word and InDesign documentation to structured DITA, centralizing content and enabling multi-channel publishing.
Global Standards
Established content creation, review, and localization standards across global teams, improving consistency and reducing time to publish.
Senior Technical Writer
Aristocrat Technologies
Gurgaon, India
Established structured documentation foundations through DITA information modeling, XML-based publishing, and enterprise taxonomy design.
Single-Source Publishing
Designed DITA information models and customized output plug-ins using XSLT and XSL-FO for PDF, HTML, and CHM formats from a single source.
Taxonomy & Validation
Developed enterprise taxonomy, controlled vocabularies, and XML validation rules for structured documentation.
Support Deflection
Reduced customer support inquiries by 15% by producing clear troubleshooting guides and FAQs.
Tools & expertise
AI & Retrieval
Information Architecture
API & Developer Docs
Agentic & Pipeline Engineering
Analytics & SEO
Web & Scripting
Certifications & awards
Professional Certifications
Prompt Engineering: How to Talk to the AIs
Designing effective prompts for LLM-driven workflows · 2025
UX Foundations: Information Architecture
User-centered IA design and navigation principles · 2021
Jenkins Essential Training
Advanced CI/CD automation and pipeline management · 2022
The Web Developer Bootcamp
Full-stack web development and modern frameworks · 2022
Awards & Recognition
Adobe Kudos! Award
Exceptional contribution to Adobe HelpX platform
ABB ERS Champion
Excellence in developer documentation and ecosystem support
Aristocrat Associate of the Month
Outstanding performance in structured content implementation
Mycom Star of the Month
Exceptional delivery and team collaboration
Impact by the numbers
Measurable results that demonstrate the business value of strategic information architecture, AI-driven tooling, and content optimization.
Improvement through standardized IA and navigation at Actian
Increase through a comprehensive SEO program
Achieved through the M&A documentation framework at Adobe
Increase via REST & GraphQL API documentation strategy
Through Jenkins DocOps pipeline automation
Supported through Adobe HelpX platform architecture