3. Data, Science & AI

Specializes in building LLM applications, RAG pipelines, and vector search systems for AI-powered feature integration.

Automates live documentation ingestion by integrating with web crawling libraries for data collection and processing.

Provides Standard Operating Procedure for epidemiological analysis using Epiverse-TRACE tools.

Provides comprehensive veterinary clinical knowledge including drug dosages, diagnosis codes, vaccine schedules, and species-specific guidance for building clinical tools.

Provides Supabase-specific analytics patterns for dashboards, including RLS-aware queries, time-series analysis, cohort analysis, and tenant-isolated aggregations.

Provides veterinary lab equipment integration patterns including HL7/FHIR message parsing, result normalization, and reference range handling.

Defines core operating directives for the Gosling Agent, including interaction style, scope management, and architectural philosophy.

Enables persistent knowledge storage and retrieval for AI agents using text embeddings, supporting cross-session context management via natural language commands.

Converts FAQ documents (XLSX, Word, PDF, TXT) into RAG-optimized formats for seamless integration into AI knowledge bases.

Builds AI agents with OpenAI Agents SDK, enabling tool registration, conversation history management, and stateless execution for conversational AI applications.

Enables development, testing, deployment, and troubleshooting of distributed data processing jobs with Ray Data, optimized for GPU and CPU resources in large-scale environments.

Optimizes veterinary clinic scheduling with AI-driven slot availability, multi-service bundling, no-show prediction, and emergency slot management.

Python framework for constructing LLM applications with chains, agents, memory, and integrations like OpenAI and Anthropic.

Manages MongoDB and PostgreSQL databases including schema design, query optimization, migrations, and performance analysis for data-driven applications.

Framework for building stateful, multi-agent LLM applications with memory, persistence, and production deployment capabilities.

Routes chat queries to optimal Azure OpenAI models based on complexity, achieving up to 60% cost reduction in AI application deployment.

Analyzes multi-agent AI workflows by tracking agent trajectories, computing metrics, and generating analytics via Azure AI Foundry cloud evaluations.

Enables vector search, RAG pipelines, and agent memory in Azure PostgreSQL via pgvector and azure_ai extensions.

Builds production-ready multi-agent systems using Microsoft Agent Framework (client-side) and Azure AI Foundry (cloud-managed), with MCP integration and orchestration patterns.

Extracts structured data from documents, images, audio, and video using Azure AI Content Understanding for invoice, receipt, and ID processing.

Provides a command-line interface to interact with external tools, APIs, and data sources via the Model Context Protocol (MCP) servers.

Offers reference schema and best practices for building stateful, intelligent AI agents based on LangGraph framework architecture.

Builds real-time AI avatars using Azure Voice Live API, GPT-Realtime, and WebRTC/WebSocket streaming for interactive applications.

Transcribes contact center calls with speaker identification using Azure AI Speech, delivering results synchronously and faster than real-time.