I build practical AI systems that connect language models with real workflows: retrieval, automation, dashboards, structured outputs, and decision support.

My focus is on Generative AI, Prompt Engineering, and AI Engineering, especially systems that use RAG, agents, APIs, and data pipelines to solve business problems. I bring a mix of software development, machine learning, analytics, and quality systems thinking, which helps me move from prototype ideas toward reliable, measurable AI applications.

I am currently open to roles in GenAI Engineering, Prompt Engineering, AI Engineering, and LLM application development.

• Generative AI: prompt design, structured outputs, LLM workflows, evaluation patterns
• RAG & Semantic Search: knowledge bases, vector databases, retrieval quality, context-aware responses
• Agentic Systems: multi-step orchestration, LangGraph-style workflows, n8n automation, approval flows
• AI Applications: Python APIs, FastAPI/Flask backends, React interfaces, analytics dashboards
• Applied ML & NLP: text classification, transformers, BERT-style QA, inference experiments
• Analytics: SQL, Pandas, KPI reporting, forecasting, business analysis

An AI-assisted inventory analytics dashboard where users can ask natural-language questions, generate governed SQL plans, view KPIs, export results, and inspect explainable outputs.

Built with: FastAPI, DuckDB, Pandas, React, TypeScript, Tailwind CSS, Chart.js, Gemini API, JWT auth, Pytest

A hands-on notebook collection exploring GPT-style text generation, Hugging Face Transformers, BERT question answering, and XLNet text classification.

Focus: prompt-conditioned generation, tokenization, hidden states, extractive QA, transformer inference workflows

A natural-language flight search app using a multi-agent flow: one agent parses travel intent into structured JSON, and another queries live flight data.

Built with: Python, Flask, Gemini 2.5 Flash, Duffel API, multi-agent architecture

A machine learning system for classifying URLs as legitimate, suspicious, or phishing using multiple classifiers.

Models: Decision Tree, Random Forest, KNN, Neural Network

A data analysis workflow for cleaning, sampling, undersampling, and preparing weather datasets for exploratory analysis and modeling.

Built with: Python, Pandas, NumPy, Matplotlib, Seaborn, Jupyter Notebook

Languages & Data

AI/ML

LLM Frameworks

Backend & Apps

Cloud & Platforms

Workflow Tools

• M.Eng. Quality Systems Engineering, Concordia University
• B.Tech. Computer Science and Engineering, Manav Rachna University
• Microsoft Azure Certified: AZ-900, AI-900, DP-900

• Production-ready RAG patterns and evaluation
• Prompt optimization for reliable structured outputs
• Agentic automation for operations and analytics workflows
• LLM application design using modern AI coding tools
• GenAI systems that are dependable, explainable, and useful beyond demos