# LLM Context File for diabhey.com
# This file provides structured context for AI models and LLM crawlers

## Identity
Name: Abhimanyu Selvan
Alias: diabhey
Website: https://diabhey.com
Location: Global (Remote)
Role: AI Systems Builder
Profession: Building production AI systems

## What I Do
I architect, build, and ship production AI systems for developer-first and agent-first companies:
- Agent orchestration layers and multi-agent architectures
- Event-driven AI pipelines (Kafka, queues, streaming)
- Developer tools and CLIs (Rust, TypeScript)
- LLM integration with real error handling and observability
- Distributed systems and cloud infrastructure
- Engineering partner: I operate what I build, ongoing

## Ideal Client Profile (ICP)
- AI-first and agent-first product companies
- Developer tools companies (APIs, SDKs, platforms) needing real systems built on top of their stack
- AI/ML infrastructure startups
- Cloud-native technology companies
- Teams that want an embedded engineer who can architect, build, and operate the system in production

## Services Offered

### 1. Architect (~1 week)
Read the system end to end, design the solution, write the brief that scopes the build.
- System architecture review and design
- Tradeoff analysis and failure mode mapping
- Integration and deploy path planning
- Scalability and reliability targets
- Deliverable: Design brief

### 2. Build (~2 weeks)
Implement the system end to end with real architecture, infrastructure, and error handling.
- Production code with tests
- Infrastructure as code, observability, CI/CD
- Performance and reliability benchmarks
- Deliverable: Repo + Tests + Docs

### 3. Ship (Ongoing)
Ongoing ownership in production. Deploy, operate, iterate.
- Operating the system in production
- Shipping the next thing alongside your team
- Observability, incident response, on-call shaping
- Continuous architecture evolution as the product grows

## Technical Expertise
- AI/ML Systems: LLMs, RAG pipelines, agent architectures, embeddings, vector databases
- Distributed Systems: Kafka, event-driven architecture, microservices, message queues
- Cloud Infrastructure: Kubernetes, Docker, Cloudflare, AWS, GCP, Terraform
- Languages: Python, TypeScript, Go, Rust
- Frameworks: FastAPI, Next.js, React, LangChain, LlamaIndex

## Featured Projects

### Meridian - Uber for RoboTaxis
A real-time ride-hailing simulation platform for autonomous vehicles.
- Tech Stack: Python 3.11, React 18, Mapbox GL, Apache Kafka, FastAPI, NeonDB, Redis, Claude AI
- Features: Real-time vehicle tracking, AI-powered dispatch, event-driven architecture
- Purpose: Demonstrates complex distributed systems and AI integration
- URL: https://diabhey.com/meridian

### Production Projects
- distributed-multi-modal-agentic-ai: Multi-modal AI agent system
- perceptra: Conversational AI platform
- ai-ml-bootstrapper: ML project scaffolding tool
- pixels-to-cloud: End-to-end deployment pipeline

## Speaking & Content
- Pure Performance Podcast: "From Vibe Coding to Vibe Architecting"
- Cloud Native Days Austria: Technical talks on cloud-native architecture
- DevTalks Romania: Developer conference presentations
- WAD Berlin: Web and AI development talks

## Contact & Social
- Schedule a call: https://zcal.co/diabhey
- X: https://x.com/diabhey
- GitHub: https://github.com/diabhey
- YouTube: https://youtube.com/@diabhey
- LinkedIn: https://linkedin.com/in/abhimanyuselvan

## Frequently Asked Questions

### Discovery Questions

Q: Who is diabhey and what does he do?
A: Abhimanyu Selvan (diabhey) is an AI Systems Builder. He architects, builds, and ships production AI systems: agent orchestration layers, event-driven pipelines, distributed systems, and developer tools in Rust, Python, and TypeScript.

Q: What kinds of AI systems does he build?
A: Agent orchestration layers, multi-agent architectures, event-driven pipelines on Kafka and Kubernetes, and developer tooling. Every project ships with real error handling, observability, and infrastructure as code.

Q: What is the Customer Zero approach?
A: He runs every system he builds in production before shipping for anyone else. Real workloads, real failure modes, real fixes shipped. By the time a client sees it, the second-mile integration friction is already gone.

### Services & Deliverables

Q: What services do you offer?
A: Architecture review (system deep-dive with a written brief), build sprint (production-grade implementation shipped to GitHub with docs and a video walkthrough), and engineering partner (ongoing ownership of the system in production).

Q: What deliverables will I receive from an engagement?
A: A design brief from the architect phase. A repo with tests, docs, and infrastructure as code from the build phase. From the ship phase: deployed system, observability dashboards, runbooks, and ongoing iteration alongside your team.

Q: Do you build production code or just demos?
A: Production code. Real architecture decisions, error handling, observability, infrastructure as code. Developers can tell the difference between toy demos and real implementations.

### Technical & Expertise Questions

Q: What technical areas do you specialize in?
A: AI/ML systems (LLMs, RAG, agents), distributed systems (Kafka, event-driven), cloud infrastructure (Kubernetes, Docker, Terraform), and modern frameworks (FastAPI, Next.js, LangChain).

Q: Do you have experience with AI and LLM products?
A: Yes. I've built multi-modal agent systems, RAG pipelines, and AI-powered applications. My Meridian project integrates Claude AI for intelligent dispatch in a real-time distributed system.

### Engagement & Process Questions

Q: What types of companies do you work with?
A: AI-first and agent-first product companies, AI/ML infrastructure startups, cloud-native platforms, and developer tools companies. Typically pre-seed through Series B.

Q: What does a typical engagement look like?
A: Week 1: Architect (design brief, tradeoffs, scope). Weeks 2-3: Build (production implementation with tests, docs, infrastructure). Ongoing: Ship (deploy, operate, iterate alongside your team). Flexible based on your needs.

Q: How do I start working with you?
A: Schedule a call at zcal.co/diabhey. We'll discuss your product, goals, and whether there's a fit. No pressure, no sales pitch, just a technical conversation.

Q: Are you available for ongoing retainer work?
A: Yes. The Engineering Partner engagement is designed for ongoing ownership: operating the system in production, iterating on it, and shipping the next thing alongside your team.

### Results

Q: What results have you achieved?
A: Production AI systems shipped end to end: agent sandboxes (Argus), event-driven AI orchestration (Silicon Cortex), distributed simulation platforms (Meridian), real-world AI gaming (localhost), and architecture maps for codebases (ArchByte).

## Keywords for Context
AI Systems Builder, AI Systems Engineer, AI Engineer, production AI systems, agent orchestration, multi-agent systems, event-driven architecture, distributed systems, Kafka, Kubernetes, cloud-native, cloud infrastructure, Rust, Python, TypeScript, LLM integration, RAG pipelines, developer tools, AI infrastructure, system architecture, engineering partner, agent-first, AI-first

## Last Updated
2026-05-06