|
|
ツ always shipping |
┌──────────────────────────────────────────────────────────────────────┐
│ MODEL DEVELOPMENT │ TRAINING & FINE-TUNING │
│ ─────────────────── │ ────────────────────── │
│ • LLM app development │ • LoRA / QLoRA fine-tuning │
│ • RAG pipeline design │ • RLHF & reward modeling │
│ • Computer vision │ • Distributed training (DDP) │
│ • Embedding models │ • Dataset curation & eval │
├──────────────────────────────────────────────────────────────────────┤
│ AGENT ORCHESTRATION │ DEPLOYMENT │
│ ──────────────────── │ ────────── │
│ • Multi-agent systems │ • OpenAI / Anthropic / CF APIs │
│ • Tool use + MCP │ • Vector DBs (pgvector, Pinecone) │
│ • LangGraph & AutoGen │ • AWS SageMaker & Lambda │
│ • Autonomous pipelines │ • Model serving (vLLM) │
└──────────────────────────────────────────────────────────────────────┘
LANGUAGES
AI / ML
FRONTEND
BACKEND
DATABASES
CLOUD & INFRASTRUCTURE
|
🤖 Jarvis — AI Portfolio Assistant Streaming LLM chat, persistent D1 history, Spotify context, RAG over resume. Llama 3.3 70B at Cloudflare edge.
|
📝 GrantAide Multi-agent grant writing — improved success rates by 85%. Writer + reviewer + optimizer agent pipeline.
|
|
🧠 Fake News Detection BERT + kNN ensemble — 92% accuracy on FNC + LIAR-PLUS. NLP multi-source verification pipeline.
|
🦁 Wildlife Conservation Bot Species detection from images + conservation chatbot. OpenAI Vision + Gradio UI.
|
|
🚗 Lane Detection System Real-time CV for autonomous driving. YOLO + NMS + OpenCV — 10 FPS, 92% accuracy.
|
🖼️ Smart Image Store Semantic search, auto-tagging, embedding-based retrieval for images.
|
|
|
always in the zone 🎧 |
