Five years of shipping SaaS that scales — and a deliberate roadmap into AI engineering. From production CRM platforms serving thousands of businesses to a structured set of production-grade AI projects.
A CRM, SMS-marketing and analytics platform distributed via the Clover App Market, with expansion into the ChatGPT ecosystem. I owned technical execution, product direction and delivery — turning a stagnating product into a growing one.
Read the full case study →An early-stage AI marketplace connecting AI-generated and human-created fashion models for small-to-medium clothing businesses. Built the technical and product foundations alongside a marketing co-founder.
View the product brief →
An AI research platform that retrieves, reasons over, and reports on your own documents and the live web — with grounded citations and human approval at every consequential step.
A standalone MCP server for searching, discovering, and navigating academic research papers through Semantic Scholar and arXiv.
AI-powered fashion photography platform for Sri Lankan fashion brands — upload clothing, pick an AI model, and get studio-quality photos in minutes.
An internal revenue intelligence platform for Clover app billing, merchant lifecycle, and growth planning, powered by Nova — an AI agent that automates financial analysis.
A CrewAI-powered product consulting board that turns a vague product idea into a validated, build-ready blueprint.
Four production-focused projects that progress from retrieval systems and agents to workflow orchestration and enterprise-scale AI architecture — each built with the same engineering principles as modern software: scalability, observability, security and cost-efficiency.
Organizations store valuable information across documents, wikis, databases and PDFs — finding accurate answers quickly is hard.
Research means gathering from many sources, validating findings, synthesizing insights and generating recommendations.
Businesses spend significant time on repetitive workflows involving multiple systems, approvals and manual decisions.
Building AI systems requires orchestration across models, agents, retrieval, evaluation, deployment and cost.
Owned technical execution, product direction and delivery. Led a full-day system-design & strategy workshop producing 20+ action items; unified engineering and marketing around shared weekly goals (~75% efficiency gain); designed the MCP-based ChatGPT App integration and shipped it in ~2 months; built analytics pipelines processing 110M+ records.
Full-stack across SMS campaign systems, AI chatbot flow-generation, and CRM/analytics. Architected an automated campaign execution engine with Kafka-based event-driven workflows; built high-performance ClickHouse analytics pipelines; led React frontend architecture; mentored 3 interns.
Stabilized a brittle legacy customer-engagement platform, retaining enterprise customers under active usage. Integrated a new telecom provider API handling thousands of messages per day and created end-to-end technical documentation.
Progressed from intern to engineer while contributing across multiple SaaS products. Helped build a core product from scratch that evolved into an enterprise-grade platform. Applied TDD with Jest, Puppeteer and Mocha within an Agile SDLC.
A full narrative of the MarketBuzz turnaround — diagnosis, alignment, architecture and crisis handling under pressure.