01 — AI engineer · six years shipping software
I design and build the agents and back-end systems that quietly do the work — for distributors, founders, and teams who'd rather see results than slide decks.
The short version
I'm Abhishek — an AI engineer who builds agent systems that businesses can actually run.
For the last six years I've been shipping software — AI and platform engineering for the past stretch of it, most recently on the agents framework behind Bluehost's AI products. Outside of that, I'm building Neev, a modular operations platform for Indian MSMEs starting with textile distribution — because the most exciting place for AI right now isn't another consumer chatbot. It's the 63 million businesses still running on WhatsApp messages and paper ledgers — 6.34 crore, per the National Sample Survey Office's 2017 Key Indicators of Unincorporated Non-Agricultural Enterprises.
My way into AI was Neural Networks: Zero to Hero, Andrej Karpathy's video course on building neural networks from scratch in code. I didn't just watch it — I built micrograd and makemore from scratch to understand what I was watching. That habit, going to the foundations rather than the abstractions, is how I work on most things. Including this site.
Five case studies, ordered by strategic weight. Each is a problem in the client's words, an approach, what shipped, and honest scope on what was and wasn't included.
Bringing AI to an industry that still runs on WhatsApp.
A modular operations platform for Indian textile distributors — built to be boring where boring matters, and quietly smart where it counts.
A retrieval stack an institution can actually trust with its own words.
Multi-tenant institutional RAG, migrating from Fastembed ONNX to Hugging Face TEI for BAAI/bge-reranker-v2-m3. Qdrant as the backbone, Ory for identity.
The foundational platform behind Bluehost's agentic AI products.
Where AI agents meet web-hosting reality — customer scale, production uptime, and real users with real bills. A major hand in maintaining and continuously improving the platform.
A fair-comparison tool for crypto cards, built like a real product.
Custody checks, provider coverage, multi-environment deploys — the boring-but-important scaffolding most crypto product sites skip.
From monorepo checks to a real-host k3s UAT, with every mock and boundary named.
A Go-service and Next.js platform taken through root-level validation, Helm deployment, and a mock-inference metering-to-billing smoke — pre-production proof, not a production or customer outcome.
First-principles notes on agent systems and AI for traditional businesses. New posts land here without a redesign.
Claude Code is where I want the work to run. GPTx keeps Claude in charge while selected GPT workers join through the same agent interface, inside the same project session.
10 min readEveryone spent this summer telling you to stop prompting and start designing loops. I agree, and I got there backwards. The brakes came first, then the map, then the gate, then a second engine.
9 min readThe last gap was not a missing rule. It was no guarantee the rules ran the same way in Claude Code and Codex, at every layer, on every project. Here is how I closed it.
9 min readThree engagement shapes. Each with what's in scope, what's not, and a realistic timeline. If something sounds close but not quite, tell me what you're actually trying to do.
A working agent system, in production, doing one thing your business needs done. Tool-use, memory, observability from day one.
For businesses that run on WhatsApp and spreadsheets. A layer that removes typing and memory load — not one that replaces judgment.
You have a prototype that works in a demo. It needs to work on a Tuesday at 3pm with 200 users. I take it the rest of the way.
This site uses the same process I'd bring to a client engagement — PRD, ADR, ROADMAP, EPM progress log, CHANGELOG, process-gate scripts. Public on the repo; linked from here.
The portfolio is itself a case study in what I build.
Email is best. Tell me what you're trying to do in plain language — the industry, who uses it, what's getting in the way. I read every one.