No matches.
↑↓ to move · ↵ to open · esc to close · j / k jump between sections
Computer Vision
Apr 2025 – May 2026
A real-time computer vision system using Haar Cascade, LBPH, and MediaPipe for workplace presence and attendance monitoring — with synthetic-data augmentation (CSDG), selective fine-tuning, adaptive loss reweighting, an 80% temporal validation rule, and robustness testing under noise, blur, and face-mask occlusion.
Agentic AI
2026
A multimodal pipeline (OpenCV, Vision Transformers, Whisper) that turns text, image, video, audio, and PDF inputs into unified embeddings, feeding a generative, agentic AI + RAG system built with LangGraph, LlamaIndex, and FAISS for planning, tool calling, memory, self-reflection, and RAGAS evaluation — deployed via FastAPI/Docker.
🚧 Work in progress — repo coming soon
AI Automation
Jun 2026 – Sep 2026
A finished, fully deployed AI job-intelligence pipeline: ranks postings by technical skill relevance with Google Gemini, applies domain-weighted matching, and generates ATS-safe tailored resumes — with document tracking and voice interaction, while a person makes every final call.
Agentic AI
2026 – Present
Optimized Research & Intelligence Operations Network — autonomously runs the full research loop for a given question:
Combines LLMs, agentic AI, multi-agent systems, RAG, MCP, autonomous coding, evaluation, optimization, and GPU/cloud execution.
🚧 Work in progress — repo coming soon
Agentic AI
2026 – Present
Adaptive Planning & Execution Engine — dynamically decides the model, agents, tools, workflow, and resources needed to solve a task. Estimates task complexity and adaptively routes between a direct LLM call, a single agent, a multi-agent system, RAG, and tools/MCP on a local or cloud model, then evaluates and optimizes before producing the final answer.
🚧 Work in progress — repo coming soon