About Karan
I am an applied AI engineer in Toronto. I take an idea a team half believes in, figure out the job it should do, and ship the agents, retrieval, and interface that make it real.
Today I am a Senior AI Innovation Engineer at Pfizer, where I build internal AI platforms and evaluate new models and vendors for teams across the company. Before that I worked with Series A and Series B startups as a design engineer, which is why I care as much about the interface as the architecture.
Outside work I founded ClawBuilders.club, organize The AI Collective in Toronto, and mentor on ADPList.

Skills
- Agentic AI and LLMs
- Multi-agent orchestration with Mastra, CopilotKit, and the Vercel AI SDK. MCP, A2A, tool calling, RAG and vector search, LLM evaluation and tracing. Claude, GPT-4.1, Gemini, NVIDIA Nemotron, and Gemma.
- Self-hosted and secure AI
- Local inference on NVIDIA DGX Spark with LM Studio, air-gapped search with SearXNG and Firecrawl, private model gateways, and sandboxed agents.
- Product engineering
- Next.js, React, TypeScript, Astro, Tauri 2, Rust, Tailwind CSS, shadcn/ui, Convex, Supabase, PostgreSQL, Node.js.
- Cloud and delivery
- AWS (Solutions Architect Associate), Azure, GCP, Cloudflare Workers, Vercel, Docker, k3s, Tailscale, GitHub Actions.
- Design and growth
- Figma to production, design systems, WCAG accessibility, A/B testing, analytics instrumentation, technical SEO.
Talks
- API Days Toronto
Panelist
The Accountability Gap - TMU Tech Week
Panelist
- University of Toronto
Guest speaker
- University of TorontoGuest talk on GenAI
Certifications
- AWS Certified Solutions Architect, Associate
- AWS Certified Cloud Practitioner
- Microsoft Certified: Azure Fundamentals
I build in public, with Toronto's AI builders
Outside work I run hands-on build nights, organize meetups, and mentor people moving into AI. It keeps me close to what builders are actually shipping.
The AI Collective, TorontoChapter organizer
Toronto chapter organizer in a worldwide community of 250K+ members. Locally, 100+ events for 500+ builders and founders.
ClawBuilders.clubFounder
A hands-on AI agent community with 600+ newsletter subscribers and five build nights, backed by Cloudflare, Rootly AI, Stan, and Ideogram.
ADPListMentor
ADPList Top 10 Growth Design Mentor milestone. 56 sessions and 2,050+ minutes of mentoring on AI and design engineering.
View the mentoring milestone
Selected build nights
Deploy AI agents with Cloudflare
S1:E5, October 1, 2026
100 builders joined Cloudflare and Rootly for an evening of agent experiments and live demos, from code review to QA and creative tools.
Read the recapAgents for image generation at Ideogram
S1:E2, with Mackenzie Dérival
A session with 20+ advanced builders, connecting agents to image generation and sharing demos of posters, thumbnails, and creative assets.
Read the recapAgent payments and content workflows
S1:E4, with Stan and Cloudflare
A hands-on workshop on agent wallets, purchase approvals, and content workflows through Stanley MCP, built alongside co-hosts and the community.
View the workshop

Talks and panels
- API Days Toronto
Panelist
The Accountability Gap - TMU Tech Week
Panelist
- University of Toronto
Guest speaker
- University of TorontoGuest talk on GenAI
Recommendations
“Whatever gap we needed to be filled, he skilled up quickly.”
“sharp, proactive, and thoughtful UX designer”
Questions people ask
What does Karan Balaji do?
Karan is an applied AI engineer in Toronto. He designs and ships agentic AI platforms, retrieval systems, and generative UI, and is currently a Senior AI Innovation Engineer at Pfizer.
What is ClawBuilders.club?
ClawBuilders.club is a hands-on AI agent community founded by Karan in Toronto, with build nights, live demos, and 600+ newsletter subscribers.
Which AI frameworks does Karan use?
Mastra, the Vercel AI SDK, CopilotKit, LangChain, and the Model Context Protocol (MCP), with models from Anthropic, OpenAI, Google, and NVIDIA, hosted in the cloud or on local GPUs.
Does Karan speak at events?
Yes. He has spoken at API Days Toronto, TMU Tech Week, and the University of Toronto, and organizes The AI Collective Toronto.
Want to work together, or talk about AI?
Send a note and I will reply within two working days. For a faster answer, book a call.