Mohammad Atif

KEYPROJECTS ConfidenceAI — LLMUncertainty & Hallucination Detection System Python 3.12, Groq API, LLaMA3, Streamlit 2025 • Built a production-grade QA web app returning LLM answers with confidence scores (0–100%), uncertainty reasoning, and missing-information surfacing — directly tackling AI hallucination in production systems. • Engineered a self-reflective reasoning pipeline using Groq API + LLaMA3; the system reasons about its own uncertainty rather than blindly returning answers — a meaningful distinction for enterprise AI reliability. • GitHub: github.com/AtifaiAslam/ConfidenceAI ContextWeaver — Production RAG System Python, PyTorch, FastAPI, Qdrant, Docker 2024 • Engineered hybrid retrieval pipeline (BGE vector + BM25 + cross-encoder reranking) achieving 40% accuracy gain over baseline; RAGAS faithfulness 0.87, P95 latency ¡2s. • Built Prometheus monitoring dashboard and full system documentation for production handover. Multilingual Sentiment Analysis API BERT, FastAPI, PostgreSQL, Docker, AWS Lambda 2024 • BERTclassifier trained on 500K reviews achieving 94% accuracy, 0.92 F1; multilingual (EN/HI/ES), handling 50,000+ daily requests at sub-100 ms latency. DAELIS —Adaptive AI Learning Platform TensorFlow, Flask, React.js, MongoDB, AWS 2023 • AI-powered personalised learning platform with collaborative-filtering engine (88% accuracy); microservices architecture maintaining 99.5% uptime on AWS. Autonomous Greywater-Fed DWC Hydroponic System A fully autonomous system that takes household greywater in one end and delivers plant-ready nutrient solution out the other. Dual ESP32 microcontrollers, 8 sensors, 3-stage treatment pipeline, and real-time cloud dashboard. Validated over a 30-day live cultivation trial with real lettuce plants. → 91% greywater recovery → 40% nutrient waste reduction → 30-day live trial → Sub-1s decisions IoT-Enabled Grey Water Purification System IoT-based grey water purification with novel floating-ball ultrasonic FOG detection, 4-stage filtration, automated pH & TDS regulation, and Flask real-time dashboard. AI/ML layer for predictive water quality classification. → pH stable 6.5–7.5 → 100ms relay switching → AI/ML predictive dosing ESP32 ×2 Flask UART pH/TDS sensors Achievements UP Government Innovation Festival Recognised · February 10–11, 2026 · Lucknow, India