3. Data, Science & AI

Intelligently merges multiple CSV/Excel files with column matching, deduplication, and conflict resolution to consolidate diverse data sources.

Optimizes LLM cost and latency via caching, model selection, batching, and prompt optimization, delivering cost breakdowns and performance recommendations.

Expert guidance for SQLite database operations using better-sqlite3 Node.js driver, covering setup, queries, transactions, migrations, and TypeScript integration.

Core NumPy operations for array creation, dtype management, shape manipulation, and memory-efficient data views.

Analyzes video content using Google's Gemini API, providing descriptions, transcriptions, timestamp references, and YouTube URL processing for extended video understanding.

Optimizes SQL and NoSQL query performance by analyzing execution plans, suggesting indexes, and rewriting queries for databases including PostgreSQL and MongoDB.

Deploys machine learning models to production using Flask, FastAPI, Docker, and cloud platforms (AWS, GCP, Azure).

Optimizes LLM deployment infrastructure with serving frameworks, quantization, and batching for low-latency inference.

Expert guidance for regression analysis, statistical modeling, and outlier detection in Python using statsmodels, scikit-learn, and PyOD.

Supports R data analysis, visualization, and statistical workflows using tidyverse, ggplot2, and reproducible research practices with R scripts and Quarto documents.

Provides foundational statistical, probability, and linear algebra tools essential for data science applications and analysis.

Provides deep learning model development and deployment using PyTorch and TensorFlow, including CNNs and transformers for production environments.

Enables efficient vectorized operations, reductions, and custom function wrapping via NumPy's universal functions (ufuncs).

Deploys large language models with high throughput using vLLM's PagedAttention and quantization for production API serving, optimizing latency and memory usage.

Enables distributed training for PyTorch models using DDP and FSDP, including multi-node setup, checkpointing, and process management via torchrun.

Automates machine learning model training using optimization algorithms on provided datasets.

Comprehensive toolkit for managing Hugging Face Spaces, including listing, space information retrieval, and runtime control for AI model deployments.

Enables fine-tuning, inference, and training of transformer models (BERT, GPT, ViT) for NLP, vision, and multimodal tasks via HuggingFace Transformers.

Measures machine learning model performance on test datasets using accuracy, precision, and recall metrics.

Optimizes embedding models and chunking strategies for semantic search and RAG applications in AI workflows.

Optimizes SQL queries for speed using EXPLAIN analysis, indexing strategies, and query rewriting across PostgreSQL, MySQL, and SQLite.

Framework for generating high-quality sentence, text, and image embeddings using pre-trained models, optimized for RAG, semantic search, and similarity tasks in production environments.

Enables efficient data manipulation, statistical analysis, and visualization using Pandas, NumPy, and Matplotlib for data science workflows.

Accesses RCSB Protein Data Bank for 3D structural data of proteins and nucleic acids, enabling search, download, and metadata retrieval for structural biology and drug discovery.