AbhiramShaji

Solutions Architect at Botpress, owning the full delivery lifecycle for enterprise AI bots across North America, Europe, the Middle East and Asia. I architect and build every bot from scratch in TypeScript using the Botpress ADK, no low-code studio. Working at a startup means each client engagement runs like an independent contract, handling spec, architecture, implementation, integration testing, and CS handoff concurrently across 17 clients and still counting. Delivered bots span autonomous customer support, e-commerce (UpsideDrinks.ca bilingual shopping assistant, Visit London itinerary chatbot), knowledge systems (HEAG/Reggie AI 2.0 with FSA Handbook KB, Back to the Roots gardening bot), WhatsApp automation (Yamaha Motor Mexico), and Gmail-integrated knowledge bots (IonOptix). Every delivery is production-deployed and tested end-to-end.

At Deloitte USI Consulting, delivered automation solutions for enterprise clients including Aflac, where I built internal tooling to automate the insurance claims revision workflow. Developed systems to process, categorize, and route claims flagged for human review, reducing manual overhead and improving throughput. Also set up Selenium automation suites for client applications, so regression passes ran without a person driving them. Stack included .NET, JavaScript, and SQL across backend services, internal dashboards, and Agile delivery cycles.

An experiment I pulled together with fellow North Island College, Vancouver Island University and University of Victoria students while we finished our degrees. AI code generation had become genuinely good by the end of 2024, and Vancouver Island is full of small businesses with no website, so we built workflows that turned out small, clean sites at the lowest price we could still deliver properly at. Money was never the goal; helping local businesses was. I led the build: Next.js, Shadcn for UI, SEO-first routing and Google Analytics. No longer accepting new clients.

Built the complete web application for ObjectStore.ca from scratch on contract for founder Jamie Starke. Handled the full stack, from architecture and database design through to frontend implementation and deployment. Delivered a production-ready platform end-to-end, no agency involved.

Langroove is a social language learning app that helps users practice languages in real-time through interactive text conversations with built-in translation support. Built using React Native (Expo) and Firebase, it integrates third-party APIs for translations and offers a modern, accessible UI for all levels.