Built a PE due diligence platform inside Bain and took it from concept to revenue in 18 months.
I build and sell B2B products from work I know firsthand.
I'm Prasad Subrahmanya. I start inside the buyer's current process, do enough manual delivery to see what repeats, then build around the repeated part. Today I run Luminik, an event pipeline platform for B2B teams.
Luminik
Event pipeline platform for B2B
Source, enrich, sequence, capture, and attribute event-sourced pipeline in the systems B2B teams already use.
Alfred OS is the agent harness I built and open-sourced to run that work: it schedules a fleet of named agents, isolates each run in its own git worktree, and opens PRs I inspect before anything ships.
I sell early because a buyer's question about event follow-up, CRM ownership, or attribution can change the product before the wrong thing gets expensive.
A few outcomes from shipped work
I have built revenue-generating software inside Bain, co-founded a venture-backed company, and shipped the first product myself before the team existed.
Co-founded Mainteny, built the first field-service product version in 3 months, and helped make the seed story concrete with a live product and customer evidence.
Engineering, product, and founder-led GTM across 4 companies. Now building Luminik.
The work before software becomes obvious
Good B2B software usually starts with a buyer already trying to get something done. The useful question is not "does this hurt?" It is "who owns this project, why now, what have they tried, and why are those options failing?" I like that stage: manual delivery, narrow first versions, sales calls that change the product, and architecture choices that keep future work cheap.
Right now I am building Luminik. Before that I built SnowOptix, Aura at Bain, and Mainteny. The settings changed. The job stayed similar: find demand in the buyer's current work, build the first useful system, sell through the workflow, and make the work clear enough that customers, teammates, and investors can trust it.
What I build and write about
Useful notes for technical founders who build, sell, and operate the first version themselves.
Solo product engineering
Ship the first useful workflow before the company has enough people to hide behind process.
Founder-led GTM
Look for the live project, the urgency, the options tried, and the gap that makes a buyer move.
AI-agent workflows
Build harnesses for digital employees: specs, repo rules, context, evals, and approval rules.
Event pipeline systems
The data model behind event ROI: sourcing, enrichment, sequencing, capture, and attribution as one record.
Start with a project someone already owns and cannot keep ignoring.
Do it yourself before freezing the product shape.
Automate the repeated work. Leave the judgment calls visible.
Let agents draft, code, check, and summarize inside explicit boundaries.
Make the model clear enough for customers, teammates, and investors.
Where I have built and shipped
Different settings, same job: understand the business, build the product, sell the early version, and keep the operating model honest.
Luminik
Founder
Building an event pipeline platform for B2B teams that need one record from source to attribution: source, enrich, sequence, capture, and attribute.
SnowOptix
Founder
A Snowflake cost optimization side project that entered Snowflake's Native App Accelerator with support from the startups team. The conversations validated the cost pain and pointed me toward events, where buyers had a live workflow and a clearer reason to move.
Bain & Company
Venture CTO
Built and scaled Aura, a PE due diligence SaaS platform, from concept to $3.6M ARR in 18 months.
Mainteny
Co-founder and CTO
Built the first field-service product version solo in 3 months, joined early customer and sales work, and helped make the seed story concrete with a live product and customer evidence.
Notes from the work
Essays on building, selling, AI-agent workflows, and the moment manual customer work becomes software.
Alfred: how I run repeated work as a solo founder
Alfred turns repeated company work into plans, summaries, PR notes, and drafts I can review. This is how I keep the work short, sourced, and in my hands.
Separating agent work from founder work
Why I moved Luminik's recurring agent work into a dedicated environment, the rules that keep it under control, and what it took three upstream bug reports to learn.
gstack, CLAUDE.md, and founder coordination
I have run my own agent harness since mid-2025: role prompts, CLAUDE.md files, a specs repo, isolated branches, and PR review. What gstack solves, where the coordination gap begins, and what stays a founder's problem.
Bring a concrete product or GTM problem
Reach out if you want to compare notes on Luminik, event pipeline, founder-led sales, or turning a painful workflow into product.