Expertise

Machine Learning (ML)

• Languages: Python, C/C++, SQL, JavaScript/TypeScript, HTML, CSS, bash
• Deep Learning libs: Pytorch, TensorFlow, JAX/Flax
• ML libs: Pandas, NumPy, Scikit-Learn, SciPy
• Gradient Boosting: XGBoost, LightGBM, CatBoost 
• Scaling/Parallelization: Dask
• AutoML (Hyperparameter Optimization): Optuna, Ray Tune, AutoKeras, auto-sklearn, Auto-PyTorch, AutoGluon,  Amazon SageMaker Autopilot,  Google AutoML, Azure AutoML, H20 AutoML
• Wrapper libs: PyCaret
• Visualization: Matplotlib, Plotly, Bokeh
• Feature selection & explainability: LIME, SHAP, Boruta
• Containerization: Docker, Kubernetes, Helm
• Cloud Computing: AWS, GCP, Azure
• Infrastructure as Code (IaC): Ansible, Terraform
• Python interactive web applications: Streamlit, Gradio, Dash
• Python (server-side) web frameworks: Django, Flask, FastAPI, BentoML 
• JavaScript (client-side) frameworks: React (Meta) / Next.js, Angular (Google), Vue, Svelte 
• Tools: Git, VSCode, Visual Studio, Jupyter, Colab,  W&B
• Coding Assistants: Cursor (Anysphere), Windsurf (Codeium), GitHub Copilot (Microsoft), Augment, Loveable, v0, Replit, Bolt [1][bolt.diy], Zencoder, Cline, Aider, OpenHands, Claude Code [1],  OpenAI Codex CLI [1], Qodo, Amazon Q, Gemini Code Assist, Trae (Bytedance), Devin, Databutton, Continue [1], Pythagora, Devika, CodeRabbit, SWE-agent, …
• Supervised Learning:
  • Classical: Linear/Logistic Regression (Ridge [L2]/Lasso [L1]/Elastic Net [L1+L2]), SVM, Naive Bayes, K-NN, Decision Trees, Random Forest, GBM
  • Neural Net based: FNN, CNN (ResNet, Inception, U-Net), RNN (LSTM, GRU) 
• Unsupervised Learning:
  • Clustering: K-Means, DBSCAN, GMM, Hierarchical Clustering
  • Dimensionality reduction: PCA, UMAP, tSNE, LDA, Autoencoder, ICA
• Advanced ML techniques:
  • Generative AI + NLP (Natural Language Processing)
    • Techniques: Attention mechanism, Transformer architecture, CLIP, Diffusion Models, PEFT, (Q)LoRA, RAG [1], Ragas, RAPTOR, SSM (Mamba, Jamba), RLHF, RLAIF, DPO [1], LoLCAT [1]
    • Libs: OpenAI, Hugging Face Transformers, LangChain, LangGraph, LlamaIndex, GraphRAG, DSPy [1], 
      LlamaHub, litellm, Guidance, NLTK, spaCy, Genism
    • LLM (Large Language Models): GPT-4 (OpenAI), Llama 3 (Meta), Claude 3 (Opus, Sonnet, Haiku)  (Anthropic), Gemini (Ultra, Pro, Nano) (Google)
    • Cloud Inference: OpenRouter, Groq, Fireworks, Together.AI, Replicate, Anyscale, Cerebras, Lepton AI, Deep Infra, Novita.AI, TGI
    • Local Inference: LM Studio, Olama, vllm, Jan.ai
    • ChatBots: ChatGPT, Gemini, Claude, Copilot, Meta AI (Llama), Grok,  Perplexity, Character.ai, Poe, Pi, You
    • Conversational AI: Azure Bot Service / Copilot Studio (Microsoft), Lex (Amazon), Dialogflow (Google), IBM Watson Assistant, Rasa
    • Text to Image: Dall-E3 (OpenAI), Midjourney, Imagen (Google), Flux (Black Forest Labs), Stable Diffusion [1] / Stable Assistant (Stability.ai), Imagine (Meta), Leonardo, Adobe Firefly, Segmind, Ideogram, Reve, ComfyUI
    • Text/Image to Video: Sora (OpenAI), Veo (Google), Dream Machine (Luma AI), Gen 3 (Runway), Kling (Kuaishou), Pika, Hailuo (MiniMax), Stable Video Diffusion, Haiper, Meta Movie Gen, Lumiere (Google), Mochi 1 (Genmo), Hunyuan (Tencent), invideo AI, Higgsfield AI
    • Text to Music:  Suno, Udio, Mureka, Riffusion, Lyria (Google)
    • Text to Speech: ElevenLabs, Chirp 3 (Google), GPT-4o mini TTS (OpenAI), Amazon Polly, Azure AI Speech
    • Agents: overview, MCP, A2A [1], Agent Protocol, ADK, OpenAI Agents SDK [1][2], OpenAI Assistants API, OpenAI Operator, CUA, OpenAI Swarm, Magentic-One (Microsoft) [1],  AG2 (Autogen), Google Agentspace [1], Anthropic Computer Use, AutoGPT, smolagents [1] (Hugging Face), CrewAI, LangGraph (LangChain Agents), LlamaIndex Agents, PydanticAI [1], OWL (Camel-AI), Swarms [1], Agno (PhiData), MetaGPT, GPT-Engineer,  OpenInterpreter,  BabyAGI,  Agency Swarm, OpenAgents, OpenAGI [1], Haystack Agent (DeepSet), Sweep, Microsoft UFO, Microsoft Semantic Kernel, Manus, …
  • RL (Reinforcement Learning)
    • Libs: Acme (DeepMind), RLlib (UC Berkeley), Stable Baselines 3 (DLR), Keras-RL2, OpenAI Baselines, garage
  • GANs (Generative Adversarial Networks)
  • GNNs (Graph Neural Networks)
  • SSL (Self Supervised Learning)