- π Building AI-powered automation systems and full-stack applications
- π§ Exploring LLM architectures, RAG pipelines, and multi-agent systems
- π€ Open to collaborate on Python/FastAPI and AI/ML open-source projects
- β‘ Fun fact: Built a GPT-2 style LLM from scratch in C++/Python with custom GEMM kernels & Flash Attention
|
Custom transformer implementation in C++/Python with hand-written GEMM kernels, Flash Attention, and AdamW optimizer. Currently scaling to 1B parameters on GCP A100 Spot VMs.
|
Multi-tenant RAG platform (SaaS) with intelligent document retrieval. Built with Next.js, FastAPI, PostgreSQL, Supabase, deployed on Railway.
|
|
Open-source EDR/XDR platform for endpoint detection and response. Real-time threat monitoring and automated incident response.
|
Community tools integrating n8n workflow automation into the Agno multi-agent framework. Bridging no-code automation with AI agents.
|
|
First-level IT support call agent powered by VAPI/ElevenLabs with HaloPSA integration. Automated voice-based ticket creation and routing.
|
Real-time trading dashboard with FinBERT sentiment analysis and LSTM price prediction models.
|
I'm always interested in discussing AI systems, automation architecture, and open-source collaboration.
β From AlsoTheZv3n

