


Built real-time inbound & outbound voice AI calling system
Designed campaign-based outbound system with intelligent prioritization
Developed ranking engine based on: o Contact priority o Campaign budget pressure o Deadline constraints
Integrated Whisper (Speech-to-Text) and ElevenLabs (Text-to-Speech) for live conversations
Implemented interruption detection & audio processing pipeline
Built scalable backend services using FastAPI with async architecture
Developed agentic AI systems using LangGraph & LangChain
Built computer vision systems (CCTV surveillance, medical imaging) using YOLO
Designed multi-agent pipelines for classification, summarization, and validation
Worked on RAG systems, conversational AI, and enterprise chatbot solutions
Delivered scalable AI solutions across e-commerce, finance, and automation domains
Tech: LangGraph, Vector Search, LLM Reasoning ● Built multi-agent RFP evaluation system using LangGraph ● Designed manager-worker architecture for proposal comparison ● Integrated vector search + LLM reasoning for decision-making ● Enabled scalable evaluation of multiple documents with high accuracy
Tech: LLMs, Multi-Lingual NLP ● Developed multi-agent product recommendation system supporting 13+ product categories ● Designed generic scalable framework for new categories ● Features included query response in <10 seconds, multi-lingual support (English, Arabic), and intelligent filtering
Tech: Whisper AI, ElevenLabs, FastAPI ● Developed end-to-end voice AI system (inbound + outbound calling) ● Implemented contact ranking system, campaign-aware prioritization, and budget & deadline pressure handling ● Used Whisper AI for STT and ElevenLabs for async TTS ● Built real-time interrupt-aware conversational system
Tech: LangGraph, APIs, Multi-Agent Architecture ● Built 5-agent pipeline using LangGraph: Categorization Agent, Summarization Agent, Relevance Agent, Numerical Extraction Agent, and Validation Agent ● Processed BSE announcements via API to extract & classify into 9 categories, generate summaries & relevance scoring, and extract numerical insights ● Implemented a final validation layer for correction, achieving 98%+ accuracy across agents
Tech: YOLO, OpenCV ● Built CCTV surveillance system with real-time detection ● Developed medical imaging models for disease detection ● Achieved high accuracy using YOLO-based architectures