Altesse Imena

Developed a responsive React application that delivers personalized space exploration experiences through interactive features including NASA astronomy picture integration, ISS tracking, and virtual planet trips.

Production-grade multi-agent system where specialized AI agents collaborate on software development tasks. Features code generation, review, testing, and security analysis with LangGraph orchestration.

Advanced voice agent with real-time interruption handling, turn-taking, and low-latency response. Includes PII guardrails and compliance filters for production deployment.

Vision-powered agent that controls computers autonomously - navigating browsers, filling forms, extracting data using multimodal AI for screen interpretation and action execution.

Comprehensive evaluation pipeline testing LLM outputs for correctness, hallucination, toxicity, and consistency. Production-ready safety framework for regulated industries.

Cutting-edge AI assistant leveraging Anthropic's Model Context Protocol (MCP) with custom FastMCP servers for dynamic tool discovery and execution.

Advanced RAG system processing documents, images, and tables. Routes queries across text chunks, image descriptions, and structured data with unified cited responses.

Custom fine-tuned LLM using QLoRA on domain-specific data (legal contracts, medical notes, or specialized code). Demonstrates advanced model training and evaluation.

Cross-modal emotion detection system fusing facial micro-expressions and vocal tone analysis for real-time emotional state inference with applications in coaching and healthcare.

Production-grade GitHub App providing automated PR reviews with security analysis, performance checks, and inline comments using LLM with structured outputs.

Optimized computer vision and NLP models deployed on constrained hardware (Raspberry Pi, mobile) with real-time inference under 50ms latency and offline capabilities.