AI training and adoption · Toronto and Waterloo
I teach teams to ship AI that holds up in production.
I'm an applied AI engineering manager. I build production LLM systems by day and teach developers how to build them the rest of the week. My programs are hands-on, evidence-based, and built around your real work.
Grand prizeGeotab Hackathon 2026
Lead instructorMakersLounge Toronto
6 years
shipping production ML and LLM systems, from edge devices to enterprise agents
50+
engineers and founders in a single agentic coding workshop, building live in the room
60%
less manual triage for a safety team, from an LLM pipeline I built and deployed
1st
globally at the Geotab Vibe Coding Hackathon 2026, with FleetShield AI
01 / The gap
Every team has AI tools now. Very few have AI in production.
Prototypes that never ship
The demo works. Nobody can prove it's good enough to launch.
Coding agents nobody trusts
Engineers use them, but the output gets rewritten or quietly ignored.
Model choice by gut feel
No evaluation harness, so every decision is a debate.
Leadership wants a plan
And the team doesn't know where AI actually pays off.
02 / Programs
Three programs, plus custom work. Each one ends with something real.
Pick a program to see what your team works through, session by session.
Half or full day · up to 50 people
Agentic Coding Workshop
For teams already using Claude Code, Cursor or Copilot who want reliable output instead of vibes.
Your team leaves with
A working project built in the session, plus a practitioner reference guide the whole team keeps using.
Pricing depends on team size and format. Contact me for a quote.
- Part 1Specs before promptsWrite specs an agent can actually execute.
- Part 2Plan mode and review loopsMake the agent think first, then check its work.
- Part 3Plugins, MCP and subagentsConnect agents to your tools and split big tasks safely.
- Part 4GuardrailsTests, diffs, and knowing when to stop the agent.
- Part 5Guided buildEveryone ships working software before they leave.
4 to 6 weeks · engineering teams
Production AI Bootcamp
For engineers building LLM features who need to get from prototype to something they can launch and defend.
Your team leaves with
One AI system shipped on your own use case, with an evaluation harness, regression gates and runbooks.
Pricing depends on team size and format. Contact me for a quote.
- Week 1How LLMs failTokens, context limits, grounding and hallucination.
- Week 2Retrieval that worksChunking, indexing and search you can debug.
- Week 3Agents and tool callingOrchestration patterns and MCP, without the chaos.
- Week 4Evaluation harnessesCorrectness, grounding, latency and time to first token.
- Week 5Tracing and debuggingFind root causes from traces, then gate every deploy.
- Week 6Launch criteriaResponsible AI controls as design inputs, not afterthoughts.
Ongoing · leaders and managers
AI Adoption Advisory
For engineering and product leaders deciding where AI fits, what to build first, and how to bring the team along.
Your team leaves with
A prioritized roadmap your team can execute, and the evidence to defend it to leadership.
Pricing depends on team size and format. Contact me for a quote.
- Step 1Use case discoveryFind where AI pays off, ranked by value and effort.
- Step 2Evidence-based model choiceDecide with evaluation results, not vendor slides.
- Step 3Build, buy or waitClear calls on what to own and what to rent.
- Step 4Risk and governanceLaunch criteria your risk team will sign off on.
- Step 5Skills and rollout planWho needs to learn what, and in what order.
Scoped to you · any team size
Custom Engagements
Need something that does not fit a package? Tell me the goal and I will shape the work around it.
Your team leaves with
A proposal built around your goal, your team and your timeline. Contact me to start the conversation.
Pricing depends on team size and format. Contact me for a quote.
- TalksConference talks and lunch and learnsPlain English sessions on AI, agents and evaluation for any audience.
- BuildHands-on build supportI work alongside your engineers on one specific AI system.
- ReviewArchitecture and evaluation reviewAn honest second opinion on an AI system before you launch it.
- BriefExecutive AI briefingsWhat is real, what is hype, and what it means for your business.
- MentorOne to one mentorshipFor engineers moving into AI and leaders building AI teams.
03 / Approach
How I teach. Practitioner first, no hype.
i.
Mental models before tools
Tools change every month. Understanding how models fail does not.
ii.
Build in the room
Every session ends with working code on your stack, not a slide.
iii.
Evaluate everything
If you can't measure it, you can't defend it to leadership or launch it.
iv.
Guardrails over vibes
Specs, tests and review loops that make agents safe to rely on.
04 / Experience
I teach what I've shipped.
Every module comes from a system I built, broke, and fixed in production.
Full profile on LinkedIn- 2024 to nowManager, Applied AI Engineering · Big Four professional services firmDesigned and launched a production LLM agent used daily by tax research teams. Built the multi-model evaluation harness behind the production model decision. Lead design and code reviews for the AI engineering pod.
- 2023 to 2024Data Scientist, LLM and Applied Analytics · Fiddlehead TechnologyLLM summarization pipeline that cut a safety team's manual triage by 60%. Telematics analytics that surfaced $1.7M in potential savings.
- 2021 to 2023Machine Learning Engineer · Quantuity Analytics5G IoT predictive maintenance built end to end, cutting client maintenance spend by 50%. Quantized models on NVIDIA Jetson. Co-inventor on a USPTO provisional patent.
- 2020 to 2021Research Associate, NLP and Computer Vision · QurHealth SolutionsMultilingual healthcare voice assistant that lifted engagement 20%. Drift monitoring and automated retraining in production.
05 / On stage and in the community
Where you might have seen me.
TMLS and MLOps WorldSpeaker, agent guardrails and evaluation
MakersLounge Toronto × TMULead instructor, agentic coding
Geotab Vibe Coding 2026Grand prize, FleetShield AI
AI Voice Agents, Waterloo2nd place
ConHacks 2026Judge
AI with VimalYouTube, Instagram and Medium
"No fluff. No generic advice. Just real, honest guidance. Walked out with more direction than I've had in months."
AI Engineer, mentee
"Thank you for believing in me, trusting my ideas, and giving me that push to stop just thinking about building something and actually build it."
Applied AI Specialist, mentee
5.0 / 5
Average rating from mentorship sessions on Topmate, with a People's Choice badge.
Read the reviews
06 / About
Hi, I'm Vimal.
I started as an electronics engineer in Chennai. Since then I've built a multilingual health assistant, edge AI on tiny devices, fleet safety analytics, and today, production LLM agents in a regulated enterprise.
Along the way I noticed something. Most teams don't need more AI tools. They need someone who has shipped this work to show them how, honestly, on their own problems. That's what I do.
B.Eng, Anna University · Graduate Certificate in AI and Data Science, Loyalist College · Microsoft Azure Data Scientist Associate · TensorFlow Developer · NVIDIA Jetson AI Specialist
Let's get your team shipping.
A 30 minute call. An honest look at where AI helps your team, and where it doesn't.
Book a discovery call or send me an email