Mahesh Chowdary Ponnaganti

• Built and deployed RAG (Retrieval-Augmented Generation) systems using LangChain and OpenAI/LLama2 models to enable real-time, context-aware student query responses.

• Fine-tuned LLMs on domain-specific educational data using LoRA, PEFT, and quantization techniques, improving model performance while reducing inference costs.

• Integrated semantic search with FAISS and ChromaDB to power vector-based document retrieval across a vast knowledge base of learning materials.

• Developed multi-agent pipelines with LangChain for personalized tutoring, auto-grading, and curriculum summarization tasks.

• Engineered prompt templates for dynamic content generation, including test creation, feedback automation, and adaptive learning support.

• Created interactive GenAI-powered dashboards and assistants using Streamlit and LangChain agents, improving user engagement and satisfaction.

• Deployed GenAI applications on AWS (S3, Lambda, SageMaker) with CI/CD automation using GitHub Actions and Docker.

• Built and maintained ML pipelines using PyTorch, Hugging Face Transformers, and scikit-learn for predictive modeling and NLP tasks.

• Implemented model evaluation techniques such as zero-shot/few-shot testing, hallucination tracking, and feedback loop integration.

• Ensured ethical AI compliance through bias audits, explainability (SHAP, LIME), and privacy-safe data practices (PII redaction).

• Monitored GenAI systems in production using tools like Prometheus, Grafana, and EvidentlyAI to track drift, latency, and accuracy.

• Collaborated cross-functionally with data scientists, engineers, and educators to align AI solutions with learning outcomes and business KPIs.

· Analyzed student enrollment trends and identified key factors influencing admission rates.

· Explored and optimized fee structures to evaluate their impact on student enrollment.

· Cleaned, preprocessed, and structured large datasets using Pandas, NumPy, and SQL for actionable insights.

· Developed predictive models using linear regression, decision trees, random forest, XGBoost, and time series forecasting to forecast student enrollments.

· Developed and deployed time series forecasting models (ARIMA and LSTM) to predict future student enrollment trends with high accuracy

· Identified trends in student joining patterns, and proposed optimal fee structures to maximize enrollment and revenue.

· Created interactive dashboards in Power BI to visualize enrollment trends, fee structure issssssmpact, and student demographics for strategic planning.

· Utilized AWS S3 and EC2 instances to store, preprocess, and analyze educational data for real-time model deployment and reporting.

· Ensured compliance with responsible AI practices by incorporating bias detection, fairness checks, and transparency reporting in model development workflows.

· Applied data governance and ethical AI principles during model training and deployment, including secure data handling, consent-based data use, and explainability techniques for model outputs.

· Recommended pricing strategies through statistical analysis and predictive modeling.

· Monitored model performance and fine-tuned predictions based on real-time data and outcomes.

· Created interactive reports and visualizations using Matplotlib and Seaborn for internal stakeholders.

· Conducted A/B testing to compare the effectiveness of various fee structures and promotional strategies.

· Collaborated with business and marketing teams to align analytics with strategic goals.

· Partnered with engineering teams to enhance data infrastructure and streamline data pipelines.

· Provided actionable insights to improve student retention, engagement, and overall user experience.

· Performed Exploratory Data Analysis (EDA) on construction and operations data to uncover trends, detect anomalies, and identify project risks.

· Developed interactive dashboards and visualizationsusing tools like Power BI / Tableau / Excel, enabling stakeholders to monitor KPIs such as cost, timelines, and resource allocation.

· Cleaned, merged, and transformed raw datasets using Python (Pandas, NumPy) and SQL, improving data quality and usability for analytics.

· Built predictive models for estimating project completion timelines using regression and time-series forecasting techniques (if applicable).

· Automated routine reporting tasks, increasing data update frequency and reducing manual work by 50%.

· Collaborated with cross-functional teams including engineers and project managers to translate business problems into data-driven solutions.

· Ensured data integrity and accuracy through validation and quality checks during data ingestion and processing.

· Created technical documentation and presented data insights to non-technical stakeholders in weekly meetings.