Invite Me to Speak, Teach, or Mentor
Conferences, hands-on workshops, engineering and MBA college guest lectures, placement-office partnerships, and corporate upskilling — on agentic AI, RAG, voice agents, and shipping production AI. Spoken at IITs and NITs. Available worldwide, online and in-person.
Topics I Speak On
Practitioner talks from systems live in production — not slideware
Agentic AI in Production
- LangGraph state machines and multi-agent orchestration
- Tool calling, guardrails, verifier patterns
- LangSmith evals and agent observability
- What "production-ready" actually means for AI
RAG That Actually Works
- Chunking, embeddings, retrieval evals
- pgvector, Supabase, hybrid search
- Semantic FAQ caching for LLM cost control
- Citation-grounded answers and trust
Voice AI with OpenAI Realtime
- STT + TTS + VAD + barge-in interruption
- Sub-1.5s p95 latency engineering on WebRTC
- Ephemeral tokens and server-side reasoning
- Stack shootout: Realtime vs Whisper vs Gemini Live
Shipping AI Like a Senior Engineer
- Distributed systems discipline applied to AI
- Cost engineering: caching, model choice, batching
- Multi-tenant AI: auth, RLS, data isolation
- Testing non-deterministic systems
For Students: A Career in AI
- What companies actually hire for in 2026
- Building portfolio projects that get noticed
- From coursework to first AI engineering offer
- Honest Q&A from someone hiring engineers
For MBAs & PMs: AI Without the Hype
- What AI products actually look like under the hood
- Cost, latency, accuracy — the real trade-offs
- Picking AI bets that ship vs ones that won't
- Working with AI engineering teams
Formats
Pick the format that fits your audience and time slot
Conference Keynote
30–60 min talk, optional Q&A. Calibrated for the technical depth of the room — engineers, leadership, or mixed.
Hands-on Workshop
Half-day to multi-day. Participants ship a working LangGraph agent / RAG pipeline / voice prototype by the end. Customised to your stack.
College Guest Lecture
60–90 min sessions for engineering or MBA cohorts. Honest, practical, no marketing — students see what real AI engineering looks like.
Placement Office Partnership
Structured prep for AI & backend roles: mock interviews, system design rounds, code reviews, AI-portfolio coaching for graduating batches.
Corporate Upskilling
1–5 day intensives for engineering teams adopting LangGraph / RAG / voice AI. Stack-aware curriculum (Java vs Python, AWS vs GCP).
Panel · Fireside · Live Online
Short-form formats: panels, fireside chats, AMAs, recorded interviews, live online masterclasses across time zones.
Who I Show Up For
Three audiences, one operating principle: be useful, not impressive
Conferences & Communities
- Topics: agentic AI in production, LangGraph patterns, RAG that doesn't hallucinate, voice AI engineering, AI cost discipline
- Format: keynotes, breakout sessions, workshops, panels — local meetups to international tracks
- What I bring: live code, real numbers from Yuvan (5 paying schools), pre-built demos, no jargon
Engineering & MBA Colleges
- Guest lectures: 60–90 min sessions on careers in AI, building shippable projects, what hiring managers actually look at
- Workshops: 1–3 day cohort programs — students ship a working AI prototype
- Placement-office partnerships: structured AI/backend interview prep, mock system design rounds, portfolio coaching
- For MBA programs: "AI without the hype" — cost, latency, accuracy trade-offs; how to pick AI bets that ship
- Free intro session for any accredited college that wants to evaluate fit before a full engagement.
Corporates & L&D Teams
- Upskilling intensives: 1–5 days, hands-on, stack-aware (Java vs Python, AWS vs GCP, your existing tooling)
- Topics: agentic AI, RAG, voice agents, FastAPI, system design, AI cost engineering, AI-assisted dev culture (Claude Code, Cursor, Copilot)
- Outcome: teams leave with shipped code, not slideware — and a roadmap they can execute on Monday
- Custom programs for L&D teams, internal AI guilds, and engineering org-wide rollouts
Past Appearances & Teaching
A teaching thread that runs from grad school to today
IITs & NITs
Invited talks at IITs and NITs across India. Topics covered span systems engineering, agentic AI, and careers in AI for the next graduating batch.
NIT Jalandhar
M.Tech in Computer Science — also served as a teaching assistant and instructor, mentoring thousands of undergraduates. The first chapter of a teaching thread that shaped everything since.
Nolan EduTech (VP, Engineering)
Designed and delivered a coding bootcamp curriculum that scaled the company to INR 30 crore (~$3.6M) in revenue in one year, with ~95% placement across cohorts in the thousands.
PhysicsWallah (Director, Engineering)
Hired and led 15 engineers; rolled out AI-assisted development culture (Copilot, automated code review) across the team — practical AI adoption inside an EdTech at scale.
Independent: AI Workshops & Talks
Ongoing workshops for engineering teams and AI cohorts on LangGraph, RAG, voice agents, and production AI quality. Roster grows continuously.
Free Course (Public)
Master SOLID Principles in Java — free Udemy course; 297 students, 4.1 rating. Distilled object-oriented design taught the way it should be taught.
For Event Organizers
Quick logistics — so you can decide fast
Travel
Open to local (anywhere in India) and international travel. Based in Indore, India. Travel and accommodation per your standard speaker policy.
Online Formats
Live virtual talks and workshops across all time zones. Quality bar is the same as in-person — same demos, same Q&A depth.
Lead Time
Conference talks: ideally 4–8 weeks. Workshops: 2–6 weeks. College guest lectures: as little as 1 week notice when slots permit.
Fees & Pro-bono
Speaking fees scale by audience size, format, and prep effort. Pro-bono slots reserved for student communities, hackathons, and accredited educational institutions.
Materials
Bio, headshot, and prior talk synopses on request. Slides and code samples shared with attendees post-event by default.
Languages
English (primary), Hindi (fluent). Sessions delivered in English unless audience prefers otherwise.