Find the right workflow
I map the current process, the people making decisions, the information they need, and the places where context gets lost.
Bangalore · AI developer and product engineer
I build AI workflows for complex work.
I help teams turn repetitive, information heavy processes into working AI products. I bring agents, retrieval, memory, tool integrations, and human review into one clear workflow that people can understand and improve.
50K+people reached by research products
100K+hours saved through research workflows
500K+people served by a commerce product
13%conversion lift through product experiments
AI workflow development
From a messy process to a product people can use.
I work across discovery, AI system design, reliability, and full stack product engineering. The model is one part of the result. The workflow around it decides whether it becomes useful.
Build the AI system
I turn that map into agents, retrieval, memory, structured outputs, and integrations that fit the way the team already works.
Make it reliable
I design evaluations, source visibility, human review, and recovery paths so the workflow can handle real inputs and honest uncertainty.
Ship the product around it
I build the interface and production system with TypeScript, Next.js, Python, Postgres, and the infrastructure needed to keep improving it.
Experience
Work across AI research, commerce, and product infrastructure.
I have led AI product development, built research systems used by more than 50,000 people, improved a commerce experience serving more than 500,000 people, and worked closely with founders from product idea through production.
Pave
Lead Engineer
Mar 2025 to present
I own technical direction across product experiments, learning tools, and applied AI workflows.
Silatus
Senior Engineer
Jun 2023 to Mar 2025
I built research and content tools used by more than 50,000 people, contributing to more than 100,000 hours saved.
CodeVyasa
Full Stack Engineer
Jun 2022 to Jun 2023
I built and maintained consumer products, including Aqualogica's commerce experience.
Assembo
Full Stack Engineer
Sep 2021 to Jun 2022
I worked on authentication and a Zoom transcription bot, then helped reduce prototype costs by 15%.
Aqualogica
Selected client work
Through CodeVyasa
I maintained a shopping experience serving more than 500,000 people. Interface experiments improved conversion by 13%.
Selected work
Products that moved from repeated work into a system.
Granveo and Insuveo explore current questions around memory and AI agents. FlutFast and KanbanCast came from earlier attempts to package recurring engineering and communication work into products.
Building now · Insurance research
Insuveo
A set of careful agents for collecting the information that insurance work depends on.
Insuveo explores agents that gather workflow data and ask for missing details. Each run starts with a plan that a person reviews, then returns the gaps and the sources behind its result.
View the buildFounder project · Personal knowledge
Granveo
A personal knowledge system that helps ideas, sources, and project memory stay connected.
I built experiments around visual graphs, semantic retrieval, source history, and agent memory. I am testing whether a system can recover useful context while showing what it knows and where that knowledge came from.
Explore the public workCommercial product · Mobile development
FlutFast
A Flutter starter kit that removes repeated setup from the first weeks of a mobile product.
I turned the plumbing I kept rebuilding into a product with authentication, onboarding, payments, analytics, backend services, and AI integrations. FlutFast taught me how packaging and distribution shape the value of engineering work.
Visit FlutFastCommercial build · Project communication
KanbanCast
A project tool for turning the work on a Kanban board into updates people can share.
I worked across the web product, browser extension, embeddable components, and a Flutter package. The experiment asked whether project progress could become useful communication without making a team duplicate its work.
Visit KanbanCastThoughts in progress
Ideas I keep testing through the work.
These are working beliefs about AI, software, learning, and the kind of engineer I want to become. Building gives me a way to test where each one is incomplete.
Intelligence should feel less scarce
A useful tool helps more people understand difficult subjects and act on what they learn. Access to an answer is only the beginning.
Context should survive the handoff
Teams lose time when the reason behind a decision disappears. Software can preserve the source, the change, and the open question.
AI will enter every useful workflow
As more work becomes digital, every document, decision, and handoff becomes a place where AI can assist. The real work is making those additions coherent.
Limits should be visible
AI earns trust when people can see its evidence, its uncertainty, and the point where someone must make the call.
Communication is part of engineering
A strong engineer can explain the problem, connect technical choices to the business, and help different people move toward the same result.
Learning should create agency
The best knowledge tools leave people more capable than they were before using them. That is the future I want to spend time building.
Writing
Essays about AI, software, and staying useful.
I write about reliable AI workflows, human judgment, connected knowledge, product distribution, and what engineers need as the work changes.
Browse all writingAI will eat software one workflow at a time
AI will spread through software as every company looks for places where a system can understand context, take action, and shorten the path to an outcome.
What I want software to give people
The tools I want to build should leave people with more context and more confidence in their own judgment.
What keeps an engineer relevant now
Technical depth still matters, while communication, learning, product judgment, and the ability to make work visible now carry more of an engineer's value.
A little context
I like work where product judgment and engineering share the same room.
I started building with early stage teams in 2021. At Pave and Silatus, I moved deeper into AI research, content generation, and decision support. CodeVyasa, Aqualogica, and Assembo taught me to connect technical changes to adoption, conversion, and operating cost.
FlutFast taught me that engineering becomes a product when someone can discover it, understand it, and use it. Granveo and Insuveo now let me explore memory, agents, and human review in domains where context matters.
I am especially interested in teams that have valuable information spread across documents, tools, and people. That is where an AI workflow can reduce repeated work while giving human judgment a better place to operate.
Available for conversations about AI products, workflow prototypes, and difficult operating problems.
Build together
Where is your team still stitching work together by hand?
If a process depends on research, documents, repeated decisions, or moving context between tools, I can help you map and prototype an AI workflow around it.