Sibghatullah

Projects

Signal IQ, Adaptive Traffic Signal Controller

January 1, 2026 – June 1, 2026

• Engineered an adaptive traffic signal controller using Python and OpenCV, implementing real-time vehicle detection algorithms that dynamically optimized signal timings - reducing average wait times by ~35% in simulated urban traffic scenarios. • Interfaced Python software with Arduino hardware, achieving 90%+ vehicle detection accuracy validated across extensive multi-condition test scenarios including low-light and high-density traffic environments.

Property Advisor Voice AI Agent

January 1, 2026 – June 1, 2026

• Built a real-time voice AI property advisor using LiveKit, Groq LLM (Llama 3.3 70B), Deepgram STT/TTS, and FastAPI - featuring persistent memory, identity resolution, and dynamic multi-criteria property search over a SQLite database. • Integrated Arize Phoenix with OpenTelemetry for full LLM observability; implemented hallucination detection and DeepEval-based faithfulness testing with concurrency benchmarking.

AI Fitness Trainer

January 1, 2026 – June 1, 2026

• Engineered a real-time AI fitness tracker using YOLOv11-Pose and Python, achieving 12+ FPS inference for exercise detection and rep counting across 10+ exercise types with high precision. • Integrated Groq LLM (Llama-3.3) and Plotly to deliver personalized workout summaries and dynamic performance analytics, with real-time form correction using pose estimation.

ResumeIQ, NLP-Powered Resume-Job Matching API

January 1, 2026 – June 1, 2026

• Developed an NLP-powered resume-job matching API using FastAPI, TF-IDF, Sentence Transformers, and spaCy, performing hybrid semantic scoring, skill gap detection, and structured alignment analysis. • Engineered a multi-stage pipeline with PDF parsing, keyword normalization, and weighted scoring (TF-IDF + embeddings + skill overlap), generating automated recruiter-ready candidate evaluation reports.

Customer Support RAG Chatbot

January 1, 2026 – June 1, 2026

• Built a production-ready retrieval-augmented generation (RAG) chatbot using LangChain, LangGraph, and Qdrant vector database, enabling context-aware semantic search with persistent memory across multi-turn conversations. • Deployed across Slack, WhatsApp, Gmail, and Notion APIs; improved contextual retrieval accuracy by 30% through optimized embedding strategies and structured prompt engineering.

CrewSaaS, Autonomous Multi-Agent Product Manager

January 1, 2026 – June 1, 2026

• Built an autonomous multi-agent SaaS planning system using CrewAI, FastAPI, React, Next.js, and PostgreSQL, with a 6-agent pipeline covering market research, feature mapping, system architecture, UX strategy, and revenue modeling. • Implemented real-time agent orchestration via Redis and WebSockets with GPT-40 as the backbone LLM, enabling end-to-end SaaS product specification generation in under 3 minutes.

About

Top Skills

Skills

Education

Experience

Projects

Certifications