I'm an AI Engineer passionate about building intelligent systems that can understand language, process information, and solve real-world problems at scale.
My primary interests lie in Machine Learning, Deep Learning, Large Language Models (LLMs), AI Agents, Retrieval-Augmented Generation (RAG), Natural Language Processing, and MLOps. I enjoy exploring how modern AI systems are designed, trained, deployed, and optimized for production environments.
Currently, I'm focused on advancing my expertise in agentic AI, distributed training, LLM infrastructure, and scalable machine learning systems. I believe that strong fundamentals, continuous learning, and hands-on experimentation are essential for building reliable AI products.
I actively work on projects involving LLM applications, AI agents, NLP systems, computer vision, and production-ready ML workflows while contributing to open-source technologies and expanding my understanding of modern AI architectures.
My long-term goal is to build impactful AI products that reach millions of users and bridge the gap between cutting-edge research and real-world applications.
| Domain | Technologies | Depth |
|---|---|---|
| ποΈ Speech AI | TTS (XTTS-v2, StyleTTS2, Bark), ASR (Whisper, wav2vec2), Voice Cloning, Speaker Embeddings | ββββββββββββ Expert |
| π€ LLM Engineering | LLaMA-3, Mistral, GPT-4o, Fine-tuning (LoRA/QLoRA), RLHF, Prompt Engineering | ββββββββββββ Advanced |
| π RAG Systems | LangChain, LlamaIndex, Qdrant, Weaviate, FAISS, Hybrid Search, Knowledge Graphs | ββββββββββββ Advanced |
| πΉοΈ Agentic AI | LangGraph, CrewAI, AutoGen, Tool Use, Planning, ReAct, Multi-Agent Coordination | ββββββββββββ Advanced |
| ποΈ Computer Vision | YOLOv9, CLIP, SAM, LLaVA, InternVL, RT-DETR, ByteTrack, Video Analytics | ββββββββββββ Proficient |
| βοΈ MLOps & Infra | FastAPI, Docker, Kubernetes, AWS SageMaker, GCP Vertex AI, Triton, ONNX, TensorRT | ββββββββββββ Advanced |
Production-grade systems - not demos, not notebooks.
| # | Project | Description | Tech | Stars |
|---|---|---|---|---|
| ποΈ | DIS-Vector | Open-source voice intelligence framework - speaker embeddings, verification & voice cloning. 97.3% accuracy | PyTorch Whisper XTTS-v2 FAISS |
β |
| π€ | Conversational AI Agent | Real-time full-duplex voice agent: ASR β LLaMA-3 β Neural TTS. <500ms latency over WebSockets | LLaMA-3 FastAPI WebSockets LangGraph |
β |
| ποΈ | Multimodal Vision RAG | CLIP + LLaVA multimodal retrieval system with Qdrant. 2.1M vectors indexed | CLIP LLaVA Qdrant YOLO |
β |
| π | Multilingual TTS Engine | 12-language production TTS. Custom prosody model. <200ms first-token latency | StyleTTS2 Bark Triton ONNX |
β |
| π | Enterprise RAG Platform | Hybrid dense + graph RAG with Neo4j. +40% relevance vs naive RAG | LangChain Neo4j Weaviate BGE-M3 |
β |
| π | Real-Time Object Tracker | YOLOv9 + ByteTrack on edge. 30fps on constrained hardware | YOLOv9 ONNX TensorRT ByteTrack |
β |
π’ COMPLETED π‘ IN PROGRESS π΅ PLANNED
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
FOUNDATION
π’ Python & ML Fundamentals
π’ Deep Learning (PyTorch / TensorFlow)
π’ Classical NLP & Transformers
π’ Computer Vision (YOLO, CLIP)
π’ Speech AI (TTS, ASR, Voice Cloning)
LLM ENGINEERING
π’ LLM API Integration (OpenAI, Anthropic)
π’ RAG Systems (LangChain, LlamaIndex)
π’ Vector Databases (Qdrant, FAISS, Weaviate)
π‘ LLM Fine-tuning (LoRA / QLoRA)
π‘ Agentic AI (LangGraph, CrewAI)
π΅ RLHF & AI Alignment
π΅ Multimodal LLMs (GPT-4o, Gemini Vision)
INFRASTRUCTURE & SCALE
π’ FastAPI + WebSocket Streaming
π’ Docker & Basic Kubernetes
π‘ Production LLM Serving (vLLM, TGI)
π‘ Model Quantization (GGUF, GPTQ, AWQ)
π΅ Distributed Training (FSDP, DeepSpeed)
π΅ MLflow + Feature Stores
π΅ CUDA Kernels & Custom Ops
ADVANCED
π΅ Multi-Agent Systems at Scale
π΅ AI Infrastructure Engineering
π΅ Research Paper Implementations
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
|
|
| Category | Tools |
|---|---|
| IDE | VS Code + Copilot, JupyterLab, Cursor |
| LLM Testing | Ollama, LM Studio, OpenAI Playground |
| Vector DB | Qdrant Cloud, Weaviate, local FAISS |
| Experiment Tracking | MLflow, Weights & Biases |
| API Development | FastAPI, Postman, ngrok |
| Containerization | Docker Desktop, Docker Compose |
| Monitoring | Prometheus + Grafana, LangSmith |
| Data | Pandas, Polars, DuckDB, Apache Arrow |
| Badge | Certification | Provider | Status |
|---|---|---|---|
| π€ | DeepLearning.AI - LLM Specialization | Coursera | β Completed |
| π | LangChain for LLM Application Development | DeepLearning.AI | β Completed |
| π§ | Hugging Face NLP Course | Hugging Face | β Completed |
| βοΈ | AWS Cloud Practitioner | Amazon | π‘ In Progress |
| π― | MLOps Specialization | Coursera | π΅ Planned |
| π¬ | Fast.ai Practical Deep Learning | fast.ai | β Completed |
I'm actively looking for AI/ML Engineer roles - full-time or contract. Let's build something impactful together.