Aadil

AI Engineer (3 years) who builds and ships production-grade end-to-end LLM and RAG systems with measurable business impact. Experienced across the full stack — from model fine-tuning (LoRA, Transformers, YOLO11) and agentic workflows to high-performance APIs (FastAPI) and responsive interfaces (React/Next.js). I take full ownership of complex AI initiatives from problem discovery to production deployment and iteration.

Software Engineer (AI/ML) at Federal Bank – FedServ (2025-09 – Present)

• Led development of a production agentic AI workflow using n8n and a LoRA fine-tuned LLM (SFT + ORPO) that enabled natural language discovery of relevant Oracle DB tables across thousands of schemas. Designed a multi-vector RAG pipeline combining table name, column metadata, and LLM-generated business description embeddings — achieving 85% success rate within 2–3 conversational turns and reducing manual discovery time from weeks to minutes.
• Engineered an end-to-end document intelligence pipeline using Llama 3.1, Qdrant, and cross-encoders, achieving 95% accuracy in extracting 21+ fields from diverse insurance documents.
• Custom-trained and productionized YOLO11s computer vision models for signature and stamp detection with parallel OpenCV inference pipelines for high-throughput verification.
• Architected a hybrid RAG pipeline combining dense vector search, BM25 retrieval, and cross-encoder reranking to significantly improve precision on domain-specific banking terminology.
• Designed a multi-heuristic confidence scoring engine with fuzzy entity resolution to automate maker-checker workflows across six weighted risk dimensions.
• Built and containerized a high-performance async FastAPI backend with mTLS security on Red Hat OpenShift, delivering enterprise-wide AI capabilities.

Associate Data Scientist at Dzitalinfo (2024-05 – 2025-08)

• Designed and deployed multiple production GenAI chatbots for client websites using GPT-4o and Gemini 1.5 Pro, integrated with custom RAG pipelines and LangChain, reducing customer support tickets by over 60%.
• Built an end-to-end meeting transcription and intelligence system using Whisper medium for transcription + diarization, followed by automated chat note generation with Gemini 1.5 Pro.
• Architected and scaled full-stack AI-powered applications (Inventory Management, Restaurant Billing, Gokarting Park Booking) using React/Next.js, FastAPI/PHP, and MongoDB/SQL.

Data Science Intern at Dzitalinfo (2023-11 – 2024-04)

• Developed a full-stack Attendance Management System with real-time reporting and React/SQL backend while starting work on production GenAI chatbots using advanced prompt engineering and Chain-of-Thought techniques.

Web Developer (Intern) at ION Technology Solutions (2023-07 – 2023-10)

• Built the complete corporate website siddachem.com from scratch using Bootstrap, HTML5, CSS3, and JavaScript for the CEO's oilfield chemicals export company.

B.Tech in Information Technology — AI & ML Specialization – GMR Institute of Technology (GMRIT) (2020-09 – 2024-04)