Pavan Sidharth J

I’m an AI developer with hands-on experience at Tychons Solutions, where I focus on Python and NLP. I built an AI-powered chatbot that actually got customers more engaged and made data retrieval way smoother by using machine learning and predictive analytics. I work well with others and always push to turn data into insights that people can actually use to improve their business.

Tychons wi: I built an AI warehouse assistant that lets people ask questions about warehouse data in plain English. Just say what you need, and the system gets it—it figures out your question, turns it into optimized SQL, and pulls live info straight from the databases. I added predictive analytics and anomaly detection, so it’s not just answering questions; it’s spotting trends, flagging anything unusual, and catching risks early. I also connected it to Microsoft Teams, so now anyone in the company can get quick answers and make decisions on the fly.

TIA Chatbot: I designed and launched TIA, a smart chatbot for Tychons’ e-commerce support. TIA chats with customers in real time, answers their questions, and guides them around the website using natural, conversational language. With TIA doing the heavy lifting, response times dropped, support got way more efficient, and the platform could handle a ton of user requests without missing a beat.

Real-Time Credit Card Fraud Detection (College Project): I created a machine learning app that catches fraudulent credit card transactions as they happen. It analyzes huge streams of transaction data, finds suspicious patterns, and keeps up with new types of fraud. Over time, it gets smarter and more accurate, which helps cut financial losses and keeps transactions safer.

Water Quality Analysis & Prediction (College Research Project): I built machine learning models to analyze water quality data and predict future trends. The system digs through old records, uncovers hidden patterns, and helps people make better decisions about water safety and managing resources. It’s a solid example of ML making a real impact in environmental monitoring.

Age & Gender Prediction Using OpenCV (College Mini Project): I combined computer vision and machine learning to estimate someone’s age and gender from their live webcam feed of their face. Using OpenCV for live frame processing and ML for the predictions, this project shows how the tech works in real situations—think security, ID checks, or access control.

Project Completion Certificate, ICT Academy partnered with IBM

• Designed and delivered enterprise-grade AI chatbots that enable users to query complex business data using natural language, eliminating the need for manual SQL queries.
• Built Natural Language to SQL pipelines that generate optimized queries in real time, improving data accessibility and reducing analyst dependency.
• Successfully integrated AI systems with Microsoft Teams, enabling organization-wide adoption and real-time decision-making through conversational interfaces.

• Developed Generative AI and Agentic AI workflows to autonomously interpret user intent, manage contextual memory, and execute multi-step reasoning tasks.
• Implemented rule-based AI agents combined with ML models to handle structured and unstructured data workflows efficiently.
• Applied prompt engineering and orchestration techniques to improve response accuracy, contextual relevance, and system reliability.

• Built predictive analytics models to forecast operational trends and identify potential risks in warehouse data.
• Implemented anomaly detection systems that proactively identify irregular patterns, enabling early issue detection and operational optimization.
• Reduced manual monitoring and reporting efforts by approximately 60%, improving speed and accuracy of business insights.

• Designed, developed, and deployed TIA, an AI-powered conversational chatbot for e-commerce customer support.
• Automated high-volume customer interactions using NLP, significantly reducing response time and improving engagement efficiency.
• Delivered scalable, real-time customer support solutions capable of handling concurrent users without performance degradation.

• Developed a real-time credit card fraud detection system using anomaly detection and classification techniques on large-scale transactional data.
• Designed ML models that adapt to evolving fraud patterns, improving detection accuracy and strengthening transaction security.

• Built machine learning models for water quality analysis and prediction, enabling proactive monitoring of environmental trends and sustainability planning.
• Extracted actionable insights from historical environmental datasets to support informed decision-making in resource management.

• Developed age and gender prediction models using OpenCV, applying image preprocessing and facial feature extraction techniques.
• Demonstrated practical applications of computer vision in security, identity analysis, and access control systems.

• Delivered end-to-end AI solutions—from problem understanding and model development to deployment and enterprise integration.
• Built scalable, production-ready Python-based backends for AI applications.
• Collaborated across teams using Git and agile workflows, contributing to clean, maintainable, and extensible codebases.