-0A0A0A?style=flat&labelColor=000){kind=icon}
> whoami
final-year student, Muthayammal Engineering College (2027 batch)
specializing in Cyber Security, currently living inside GenAI & Voice AI systems
> cat current_focus.log
[learning] multi-agent orchestration patterns (hub-and-spoke, deterministic execution layers)
[learning] RAG internals — hybrid search, reranking, chunking strategies that don't fall apart
[learning] voice pipelines — STT/TTS tradeoffs, latency budgets, turn-taking, cost modeling
[learning] Azure AI certification track — AI-900 done, AI-103 in progress
[always] asking "where does this break in production" before "does this demo work"|
Agentic Systems
Voice AI
|
Retrieval & Knowledge Systems
|
I don't chase certifications for the badge — I chase them because they force me to learn the parts I'd otherwise skip.
open this if you want the unusually specific list
| Layer | What I'm learning inside it |
|---|---|
| Orchestration | LangGraph-style stateful multi-agent graphs |
| Retrieval | Qdrant, hybrid ranking, semantic vs. lexical tradeoffs |
| Voice | Pipecat, real-time audio streaming, WebRTC concepts |
| Cloud/AI Platform | Azure AI Foundry, agent services |
| Security | Secure secrets handling, access boundaries in automated systems |
| Infra-adjacent | Remote execution (WinRM/SSH), observability basics |
I don't trust a concept until I've broken it, watched the logs, and figured out why it broke. Theory first, then I go poke the real system until it stops lying to me.
✨ if you're learning any of the same things — RAG, voice AI, agent orchestration — I'd genuinely like to hear how you're approaching it. Open an issue, say hi.