I build the whole system.
Fifteen years in distributed systems and cloud-native backends, now spent shipping complete production platforms solo — architecture, data pipelines, LLM infrastructure, deploy. Three of them are running tonight.
Three ways this is usually bought
AI product engineering
LLM pipelines that do real work rather than demos.
Retrieval, synthesis and evaluation; agent tooling and MCP servers; eval harnesses that tell you whether a prompt change actually helped. The interesting engineering is almost never the model call — it is everything either side of it.
Proved by prompteval, mcp-multi-db and the VivaTrades synthesis pipeline.
Solo end-to-end delivery
Architecture through production and beyond, by one person.
For a product that needs to exist rather than a team that needs to grow. No coordination overhead, no handoff between design, backend and infrastructure, because they are the same conversation. It ends with a system someone else can operate.
Proved by VeoCabs — built end to end, and running a real business since.
AI-assisted dev enablement
Claude Code and agent workflows for teams that already ship.
How to get real leverage from AI-assisted development without the review burden eating the gains: where to put the guardrails, what to automate, what to keep human, and how to tell the difference before it costs you a release.
Proved by the fact that three production platforms exist at all.
Three platforms, built end to end
VeoCabs
Solo end-to-end deliveryAn intercity cab lead-generation platform across 150+ Indian cities: instant fare estimates from Google Routes, enquiries pushed to Telegram, and a CRM the business runs on daily. Built and still maintained by me, and run day to day by the business itself.
150+ cities served37 curated route pages₹0 monthly infrastructure costRead the case study
VivaTrades
AI product engineeringMakes a stock decision take two minutes instead of an evening. Plain-English analysis of 500 NSE stocks, refreshed on a schedule, plus free strategy backtesting. No paywall, no signup.
500 nse stocks, in plain english20yr free strategy backtesting₹0 cost to use · no signupRead the case study
Shubh Bhagya
AI product engineeringGives away what other astrology sites charge for: kundli, matching, dashas, numerology, muhurat and dosha calculators, and unlimited questions answered — computed properly, cited, and free with no signup.
₹0 for everything, no signup76 cited explainers13s→1s homepage, 32 storage ops → 0Read the case study
Tools that came out of the work
- mcp-multi-dbTypeScript
Read-only MCP server over PostgreSQL, MySQL and SQLite. Safe by default, because a query tool with write access is an incident waiting for a prompt.
- promptevalPython
LLM eval framework with first-class token-cost tracking. Answers the question everyone skips: did that prompt change save money without losing quality?
- ai-support-botTypeScript
Upload documents, get a RAG-backed support assistant you can embed on any site as a widget.
- multi-agent-blogTypeScript
Cooperating agents gather context, draft, critique, then revise against the critique. A human still approves before anything publishes.
- mcp-zerodhaJavaScript
Read-only MCP server putting broker data — prices, portfolio, history — in front of AI agents.
Notes from building the whole thing
- 2 September 2026 · 6 min
What actually goes wrong with a team of agents
Running architect, engineer, QA and dev-lead agents across three production repos. The interesting failures were not bad code — they were agents manufacturing their own authority.
- 1 September 2026 · 6 min
When the API places real orders
A trading daemon runs unattended for six hours a day on a real brokerage account. The engineering that matters is not the strategy — it is what happens when something restarts at the wrong moment.
- 31 August 2026 · 7 min
What a prompt change actually costs
Everyone asks whether the cheaper prompt hurt quality, and almost nobody answers with evidence. Most of the difficulty turns out to be statistical, not architectural.
What are you building?
A short description of the problem is enough to start. If it is not something I should take on, I will say so and point you somewhere better.