Aditya Mahale

I’m an AI Engineer with hands-on experience building LLM and machine learning applications that solve real business problems. I’ve developed enterprise knowledge assistants, AI-powered SQL generation systems, and backend APIs, working across the full AI development lifecycle—from data preparation to prompt engineering, deployment, and continuous optimization.

In my Software Intern role at Geany Softech Pvt. Ltd., I designed and developed a LangGraph-based enterprise knowledge assistant by scraping and structuring company website data for natural language retrieval using FastAPI and LLMs. I improved response relevance through prompt engineering, prompt evaluation, and context optimization, and engineered FastAPI REST APIs with MySQL logging pipelines to track user queries, responses, and token usage.

I built real-world projects like NL2SQL, an AI-powered natural language to SQL query generator using Google Gemini with schema-aware prompt engineering, validation, and execution safeguards. I also created ChurnSense, an end-to-end churn prediction pipeline on 7,043 telecom customer records, using XGBoost and hyperparameter tuning, with ROC-AUC of 0.85, plus SHAP explainability to identify churn drivers and support retention strategies.

I’m motivated by measurable outcomes and production readiness, especially the parts many teams struggle with—LLM evaluation, context optimization, and reliable integrations. I’ve also filed a patent for an ML-Based Personalized Nutrition System (2025) and placed as runner-up in a Datathon (2025), reinforcing my focus on building practical AI that teams can use.