--:--:--
my name is
SHIVA BHASKAR
about me!
I'm Shiva,
a high agency product engineer in Bangalore who gets unreasonably interested in making AI do useful, boring, important work.
explore my work!
FeaturedWorks
Apr 21, 2026
Pulse Market Intelligence
An AI-driven stock market app that surfaces the why behind price movements — complete with delayed attribution, portfolio personalization, and a real-time research chatbot.
View projectMar 1, 2026
Zapigo
From a receiver-chosen gifting platform to a full digital invites and guest management product — three years as a founding team member across design, engineering, AI pipelines, and brand partnerships.
View project
Mar 2026
Back Office Intelligence Pilot
A lower mid-market PE firm with 8+ portcos running different ERPs turned a 15-day month-end fire drill into an auditable, agent-driven process.
View project
Nov 2025
Salesbuddy
Field sales reps live in WhatsApp. So we built there instead of asking them to change — store check-ins, order execution, escalation alerts, all without leaving the chat.
View project
skills!
AI is the main thing. Product thinking is what frames it. Design and code are how it ships.
- Agents & toolsPrompt systemsWorkflow designEvalsAutomation ops
- StrategyDiscovery0 → 1PrioritizationRoadmaps
- UX systemsUI designPrototypingStorytellingVisual direction
- React / Next.jsFrontend systemsAI integrationsAPIsShipping
- MentoringWorkshopsWritingFeedback loopsCommunity
also:Hardware Tech · ESG · Illustrations · Research
Teaching AI agents forced me to actually understand them.
Guest lecturer at Sahayadri College of Engineering & Management, Aug–Dec 2025. Taught agentic AI systems to third-year CS students - orchestration, tool use, observability, the parts that don't work the way the papers say they do. Some of those lectures became Substack posts because the effort put in to create content was really high and I wanted others to read and learn as well.
- Read on Substack
Agentic AI, Lesson 1
A grounding pass through the language, promise, and reality of agentic AI before the architecture gets more complicated.
- Read on Substack
Agentic AI, Lesson 2: Agents & Workflows 101
Where workflows stop, where agents start, and why the difference matters once systems need choices and guardrails.
- Read on Substack
Agentic AI, Lesson 3: Parts of AI Agents
A map of the moving pieces inside agent systems: instructions, models, tools, memory, state, and runtime scaffolding.
- Read on Substack
Agentic AI, Lesson 4: Tools & Memory
A practical look at tool use and memory design, including the parts that help reliability and the parts that quietly break it.
- Read on Substack
Agentic AI, Lesson 5: Getting Started
How to start small, choose the right first build, and move from theory to a narrow system that can actually be tested.
- Read on Substack
Agentic AI, Lesson 6: Workflows + AI
Workflow patterns, orchestration tradeoffs, and the boring reliability work that makes AI systems useful outside demos.
More Lectures in Progress
The next batch goes deeper into the tooling and evaluation layers that make agent systems easier to build, inspect, and trust.
LangChainLangGraphEvalsMCPs
Let's Connect
I'm most energized by projects where I can dig into a gnarly problem, collaborate closely, and ship something that genuinely improves someone's day.
