3. Data, Science & AI

Builds AI features with Vercel AI SDK, enabling streaming chat and structured outputs via useChat, streamText, and generateObject.

Manages LLM context windows, handles compaction, and persists state for long tasks and agent coordination.

Coordinates multiple AI agents for task delegation and specialist assignment in multi-agent workflows.

Identifies recurring patterns in data and work contexts to enable data-driven insights and decision-making.

Provides advanced data science capabilities including statistical modeling, A/B testing, causal inference, and predictive analytics using Python, R, and SQL for data-driven decision making.

Builds scalable data pipelines, ETL/ELT systems, and data infrastructure using modern tools like Spark, Airflow, and dbt for data modeling and governance.

Provides advanced computer vision capabilities including object detection, segmentation, and model training using PyTorch, OpenCV, and vision transformers for production deployment.

Expert prompt engineering for LLM optimization, RAG, agent design, and AI product development.

Assists in designing, querying, optimizing, and administering MongoDB and PostgreSQL databases including schema design and performance tuning.

Expert guidance for productionizing ML models, implementing MLOps, and building scalable ML systems with LLM integration, fine-tuning, and RAG.

Enables real-time voice AI applications with Azure AI Voice Live SDK, supporting bidirectional audio, speech-to-speech translation, and voice assistant features in Python.

Declarative hyperparameter management tool for ML/AI experiments, enabling type-hinted CLI parameters, configuration namespaces, and hyperparameter sweeps.

Stress-tests forecasts by assuming failure to identify tail risks and unknown unknowns, especially at extreme confidence levels (over 80% or under 20%).

Combines vector and keyword search to enhance retrieval accuracy in RAG systems and search engines.

AI-powered image analysis for describing scientific figures, plots, experimental setups, and extracting information from visual content.

Dynamic system for creating, assigning personalities, and orchestrating custom AI agents with voice mapping and parallel execution capabilities.

Enables multiple AI agents to engage in structured debates, providing diverse perspectives on a topic.

Enables progressive URL scraping using Bright Data for efficient web data extraction and collection.

Generates dynamic AI prompts using templates, standards, and patterns for optimized model interactions and programmatic composition.

Enables real-time bidirectional voice communication with Azure AI for voice assistants, chatbots, and speech translation via Python SDK.

Provides expert guidance on Polars DataFrame library for efficient data processing, including time series operations and performance optimization.

Analyzes Federal Election Commission (FEC) campaign finance data, including contributions, disbursements, and committee reports, for political finance insights.

Provides real-time and forecasted electricity consumption, demand data, and UECM information for Turkey. Triggers on Turkish electricity terms.

Queries Turkish electricity market data including Day-Ahead (GÖP), Intraday (GİP), and Balancing Power (DGP) markets, order books, and bilateral agreements.