I automate the B2B grind.
I sell before I build.
I'm Prasad Subrahmanya.
I build Luminik, an event pipeline platform for B2B marketing teams.
The work converts to revenue.
Pipeline sourced for customers, ARR built from zero, and products people pay for.
This is just the public footprint. Most product and client work ships in private repos.
Pipeline sourced for customers at Luminik, connecting B2B event spend to real revenue.
ARR built from zero in 15 months as venture CTO, used by top private equity and growth firms.
Lead-to-opportunity from 43,000 registered attendees to 1,840 ICP matches, up from 1.3% the year before, roughly 6x.
Founded Luminik, SnowOptix, and Mainteny. Hands-on from first commit to scaled team.
Six rules I build by.
Learned from putting agents in front of real customers, where demos do not count.
Get close to the problem
I work next to the people who have the problem, so the product comes from what they do day to day.
Treat agents as systems
An agent is more than a prompt. It needs planning, isolation, review, and a way to recover when it goes wrong. Alfred runs a fleet of coding agents on those rules.
Earn trust before production
Demos are easy. I spend the real effort on the cases that break: evals, a judge I have calibrated, and adversarial tests, so I know how a system behaves before customers meet it.
Expect failure
Long-running AI work fails partway through. I build so a run can crash, retry, and pick up again without losing data or repeating work.
Measure what matters
I tie systems to the number the business cares about: pipeline, revenue, and cost. If I cannot see the effect, I am careful about shipping.
Stay hands-on
I write code, review it, and read what the system produces in the wild. This work is hard to lead from a distance.
The stack I'm hands-on with.
Where I build, by area. Real tools, running in production.
Agentic systems
Multi-agent orchestration, autonomous agent fleets (Alfred), LangGraph, MCP, tool use and planning.
RAG & memory
Vector DBs, embeddings, hybrid and semantic search, context engineering, agent memory.
Evals & guardrails
LLM-as-judge, eval harnesses, benchmarking, red-teaming, adversarial tests, safety gates.
LLMOps & reliability
Tracing (Langfuse), OpenTelemetry, cost and latency control, failure handling, durable execution.
Applied ML & model work
RL and reward design, prompt and context engineering, fine-tuning, multi-model routing (Vertex, OpenAI, Anthropic).
Distributed systems
Microservices and SOA, event-driven and async services (asyncio, aiohttp), FastAPI, Django, Spring Boot + Kotlin, Kubernetes, Kafka.
Full-stack & mobile
Next.js, React, TypeScript, React Native and Expo, Tauri, end to end from data model to UI.
Cloud & data platform
AWS, Terraform, Snowflake, data pipelines and warehousing, Postgres, CI/CD.
Events, CRM, and product data come in as the raw input.
Agents plan, act, and hand the work off to each other.
Judges, guardrails, and error bars prove the behavior first.
Every run can retry, resume, and recover without losing work.
It all ties back to pipeline, ARR, and resolution.
I build the first version, then the team.
Across four companies, I've owned the product, the architecture, and the people who build it.
Building a multi-agent platform that connects B2B event spend to measurable pipeline: attendee extraction, enrichment, ICP matching, outbound, CRM sync, and revenue attribution. Powering GTM teams at flagship events.
A Snowflake cost-optimization tool. It was used by one of the global top-3 consulting firms, and the conversations while building it surfaced the bigger problem that became Luminik.
Took a workforce-analytics SaaS platform, used heavily in PE and growth-equity due diligence, from concept to $3.6M ARR in 15 months. Owned product, architecture, and the engineering team, and built the data infrastructure: a medallion warehouse on Snowflake with a Cube.js semantic layer, plus Lightcast and BLS taxonomies for comparable workforce data.
Field-service management SaaS for maintenance companies across Europe. Built and launched the MVP solo in 3 months, the work helped raise a $2.7M seed, and the team scaled to about 15 across five countries.
Engineering Manager at Quantumrock (BaFin-regulated algorithmic trading, Munich). Senior Software Engineer at Wirecard (PCI-DSS card processing at scale, Munich). Software Engineer at CommerceIQ (retail price intelligence, 100M+ SKUs/day). Platform Engineer at BlueJeans, later Verizon (video platform, recording v1.0, Bangalore).
Primers on what I actually build.
Field notes and first-principles primers on the systems behind the work.
Let's talk.
If you're building something hard in AI, or hiring someone who has, I'd like to hear about it.