Krishna Krishna

Data Scientist with 5 years of experience delivering end-to-end machine learning, NLP, and generative AI solutions across energy, higher education, real estate technology, and retail. Career progression spans full-stack development, data engineering, and analytics into predictive modeling, transformer-based NLP, and production RAG systems on Azure. Proven record of turning fragmented enterprise data into forecasting models, explainable ML risk indicators, document intelligence pipelines, and AI copilots that reduce manual effort and drive data-informed decisions.

Data Scientist – Generative AI | Generative AI University Copilot & Enrollment Intelligence Platform - Cleveland State University - Cleveland, OH

(2025-11)

• Developing a RAG-based AI copilot using Azure OpenAI, Azure AI Search, LangChain, and LangGraph that retrieves university policies, enrollment information, and academic procedures scattered across systems and documents, generating grounded, context-aware answers for staff and students.
• Engineered retrieval pipelines with vector search, source grounding, and citation of institutional documents to minimize hallucinations and ensure responses are traceable to authoritative sources.
• Integrated enrollment and student-service analytics into the platform, surfacing trends that support advising, capacity planning, and service improvements.
• Implemented role-based access controls, monitoring with Application Insights, and user feedback loops to continuously evaluate answer quality and safely operate the copilot in production on Azure App Service.
• Connected the solution to PeopleSoft and Stellic via REST APIs and built React/TypeScript interfaces, reducing time staff and students spend searching institutional resources by 60%.

Data Scientist (AI/NLP) | Document Intelligence & Enterprise Risk Analytics Platform - AIMNEXT

(2025-02 - 2025-10)

Fortune 500 Retail Client

• Built an NLP document-processing pipeline using Transformers and BERT to extract, classify, and standardize key business concepts from large volumes of unstructured enterprise documents, replacing slow, inconsistent manual review.
• Developed machine learning models supporting risk analytics and prioritization, with human-in-the-loop review and validation workflows to keep experts in control of high-impact decisions.
• Implemented RAG and vector search on Azure AI / Azure OpenAI to enable semantic retrieval over the standardized document corpus.
• Delivered secure FastAPI services and React front ends, containerized with Docker, with MLflow experiment tracking and model monitoring for enterprise-grade deployment.
• Reduced manual document review effort by 70% and improved consistency of extracted terminology across teams.

Data Scientist | Property Revenue Forecasting & Portfolio Performance Analytics Platform - MRI Software - Solon, OH

(2024-05 - 2024-12)

• Designed scalable ETL pipelines (SQL Server, SSIS, Azure) consolidating property and financial data distributed across multiple enterprise modules into analytics-ready datasets.
• Developed time-series forecasting and predictive models for revenue forecasting and portfolio performance analytics, replacing manual reporting with repeatable, data-driven projections and improving forecast accuracy by 25%.
• Built Power BI dashboards and full-stack services (C#, .NET, React, TypeScript, REST APIs) delivering performance trends, financial insights, and forecast results directly to business teams.
• Contributed to CI/CD-based release processes, improving reliability and speed of model and dashboard deployments.

Data Analyst | Enrollment Forecasting & Student Success Prediction - University of Cincinnati – DT Solutions - Cincinnati, OH

(2023-02 - 2024-04)

• Built ETL pipelines combining admissions, demographic, academic, and historical enrollment data from separate systems into unified datasets for institutional analytics.
• Performed exploratory data analysis and developed time-series and regression forecasting models to predict enrollment trends, supporting enrollment planning and resource allocation.
• Developed classification models (Scikit-learn, XGBoost) using academic trends, enrollment patterns, course progress, and engagement indicators to identify students at risk of academic difficulty or attrition.
• Created explainable risk indicators with SHAP and interactive Power BI/Tableau dashboards, enabling authorized university teams to spot retention patterns early and intervene with timely support, improving early identification of at-risk students by 35%.
• Shifted reporting from reactive manual reports to proactive, model-driven insights consumed via dashboards and REST APIs on Azure.

Data Engineer / Full-Stack Developer | Enterprise Data Management & Text Analytics Platform - Tata Consultancy Services (TCS)

(2021-05 - 2022-12)

Fortune 500 Energy Client

• Developed a centralized data-processing platform to clean, standardize, validate, and categorize enterprise data previously stored across multiple systems with inconsistent terminology, duplicate records, and manual validation.
• Built NLP-based text extraction (Python, spaCy, Scikit-learn) to identify and standardize key terms from operational records, improving data consistency for downstream analytics.
• Automated data-quality checks and ETL workflows in Python and SQL, reducing manual validation effort by 50% and accelerating time-to-insight for analytics teams.
• Integrated analytics into a full-stack application (REST APIs, Java/.NET, React) with data visualizations used by business and operational teams.

Master of Science - Information Technology - University of Cincinnati (2024-05)

Bachelor of Technology - Information Technology - RVR & JC College of Engineering (2021-04)