Eman Ali

Built an AI-powered restaurant auditing platform scoring businesses across 8 dimensions using 10+ external services, generating audit reports and personalised sales presentations. Developed async backend workflows with retries, timeouts, fallback providers, and Redis-backed queues with shared rate limiting.

Improved 'Best Deals' and 'New Deals' recommendations in production, serving 4M+ users. Designed a fault-tolerant inference architecture on GCP and built a dynamic discounting engine, fine-tuned Qwen 0.8B, and improved search for 150K+ products with BM25.

Built a full-stack multi-agent recruitment platform with state-based, goal-driven agents for job matching and resume tailoring. Added a hallucination-removal loop using LLM-as-a-Judge and implemented RAG with Qdrant.

Contributed to an agentic AI chatbot for natural-language database queries over restaurant data. Evaluated a multi-LLM pipeline improving intent accuracy from 92% to 98% and table selection from 58% to 90%, cutting system cost by 83% with Redis caching.

Cut AI voicebot response time from 2000 ms to under 600 ms using multithreading and async processing. Built a real-time AI DRS for LBW detection using Kalman Filters, SORT, and YOLO + Deformable DETR.

Built and fine-tuned a Roman Urdu chatbot using one open-source and three custom datasets. Applied prompt engineering and multilingual fine-tuning techniques to improve LLM responses.

Grade: 3.75/4.0

Activities and societies: 5× Dean’s List, 2× Rector’s List, Silver Medal (Sem 6), Gold Medal (Sem 7)

Bachelor of Science in Data Science with Magna Cum Laude honors, Dean's List, and multiple medals.