I build intelligent systems that combine LLMs, agentic AI, RAG pipelines, real-time processing, and multimodal interfaces.
From deploying production-grade Azure AI Search RAG stacks to creating LLM-powered robotic agents with real-time audio I/O, I love solving complex problems through practical, scalable engineering.
- 🔭 Data Scientist @ Ghaia.ai — architecting LLM agents, NL2SQL pipelines, RAG systems, and enterprise orchestration frameworks.
- 🧠 Currently deep-diving into Liquid Neural Networks, multi-agent systems, and advanced RUL modeling.
- 🤝 Open to collaborating on Agentic AI, Robotics + LLMs, Predictive Maintenance, or RAG tooling.
- 💬 Ask me about LLMs, AutoGen, RAG design, MLOps, Time-Series Forecasting, Deep Learning, FastAPI, Docker/K8s.
- 📫 Reach me at devparker48@gmail.com
- ⚡ Fun: I love building end-to-end ML systems that make real-world decisions — whether it's a robot shaking hands or a model predicting machine failures.
Real-time conversational agent with audio I/O, gesture execution, and SDK motion control.
Includes speech understanding + action execution + memory for fluid human–robot interaction.
Built a multi-agent RAG framework with:
- Azure AI Search + SharePoint/Blob sync
- DocIntel parsing (PDF, PPT, DOCX, XLSX, Images)
- NL2SQL querying
- Short-term memory & chain-of-thought orchestration
Created LSTM/TFT models for estimating Remaining Useful Life of vehicle components using voltage, temperature & sensor telemetry. Achieved high fault-classification accuracy on U.S. fleet logs.
End-to-end forecasting system using TFT + LNN, integrating micro/macro indicators, backtesting, and MLflow tracking.
LLM+Vision based system for seniors to understand medicines through image upload & voice.
Built on Gemini, Pinecone, LightRAG, and optimized with GPU acceleration + Streamlit UI.
(Projects — page 2) :contentReference[oaicite:6]{index=6}
A racing controller using computer vision & physical interaction via a tracked board.
Python · JavaScript · WebSocket · FastAPI · Flask · Docker · Kubernetes · MLflow
Azure · GCP · CosmosDB · MySQL · MongoDB
Transformers · PyTorch · TensorFlow · HuggingFace
RAG Architectures · AutoGen · Embeddings · TFT · LSTM
Volatility Modeling · RUL Estimation
AutoGen · FastMCP · NL2SQL · Azure AI Search
Doc-Intel · Multi-model embeddings · Context-aware memory
| Stats | Streak | Languages |
|---|---|---|
