Data Science
Predictive modeling, forecasting, segmentation, churn analysis, evaluation, dashboards, and insight communication.
Open to Work
Turning data into models, systems, and decisions.
I'm Neehanth Reddy, a Data Scientist and AI/ML builder with 2+ years of applied experience across predictive modeling, forecasting, machine learning pipelines, biomedical AI research, and production-style AI applications.
About
Data science first, engineering discipline underneath.
I position my work around business-facing data science: understanding the problem, building measurable models, validating results, and packaging outputs so stakeholders or systems can actually use them.
Machine Learning
Reproducible training workflows, model APIs, Dockerized delivery, CI/CD, DVC, and AWS deployment patterns.
AI Engineering
RAG pipelines, retrieval benchmarking, structured LLM workflows, agentic systems, and tool-aware application design.
Capabilities
Practical stack for data.
Data Science & Analytics
Machine Learning
Data Engineering
AI Engineering
Selected Projects
Proof of work, ordered for Data Scientist roles.
Forecast Accuracy
18% RMSE ↓
Data ScienceGitHub
E-Commerce Analytics Pipeline
Built segmentation, association-rule mining, and SARIMA forecasting on 10K+ transactions to support marketing and supply-chain decisions.
- Reduced forecasting RMSE by 18% versus baseline.
- Delivered stakeholder-ready Power BI reporting.
- Connected customer segments to product-bundling opportunities.
Pythonpandasscikit-learnSARIMAPower BI
Churn Model
91%
validation accuracy · 0.81 F1
Data ScienceGitHub
Telecom Churn Prediction
Created a churn prediction pipeline for 7,000+ customer records with feature engineering, class balancing, and model comparison.
- Improved minority-class detection over logistic regression baseline.
- Used resampling techniques to address class imbalance.
- Produced ranked predictors for retention strategy.
Pythonscikit-learnSMOTEANNF1 Score
27%Churn
$456KMonthly Rev.
$139KChurn-related loss
AnalyticsGitHub
Customer Churn Analysis Dashboard
Built an interactive Power BI dashboard on 7,043 telecom customers to quantify churn rate, revenue exposure, and high-risk segments.
- Connected churn behavior to revenue impact.
- Combined Python preprocessing with SQL KPI extraction.
- Prioritized retention actions for business users.
Power BISQLPythonKPI Reporting
106-subject baseline
ML ResearchGitHub
EEG Signal Processing & Classification
Built reproducible EEG workflows for sleep spindle detection and motor imagery classification with leakage-aware evaluation.
- Audited 109 subjects and established a clean 106-subject baseline.
- Used LOSO validation with CSP + LDA and CSP + linear SVM.
- Generated PSD visuals and confusion matrices for reporting.
PythonMNESciPyscikit-learnYASA
94%Validation
ML EngineeringGitHub
Poultry Disease Classification System
Developed a VGG16 transfer-learning classifier for four poultry disease classes and packaged it as a deployment-ready inference API.
- Versioned data and experiments with DVC.
- Served predictions through FastAPI and Docker.
- Automated deployment through GitHub Actions, AWS ECR, and EC2.
TensorFlowVGG16DVCFastAPIAWS
≥0.998 recall
7,641 chunks · 76 queries
AI EngineeringGitHub
GenAI Research Assistant
Built a RAG-based research assistant over arXiv papers with retrieval-backend benchmarking and end-to-end quality evaluation.
- Compared FAISS-HNSW, Annoy, and Flat Index retrieval.
- Evaluated latency, recall, memory, and RAGAS metrics.
- Built a Streamlit semantic Q&A interface.
LangChainFAISSAnnoyChromaDBRAGAS
Resume
JD
Fit Analysis
Structured LLM Evaluation
LangGraph · Pydantic · LangSmith
AI Engineering
GitHub
Agentic LLM Evaluation Workflow for Resume Screening
Built a LangGraph-based workflow that evaluates resume-job fit through structured extraction, gap analysis, deterministic score calibration, and recommendation generation.
- Extracted structured resume and job-description data in parallel using typed LLM outputs.
- Used Pydantic validation to make scoring inputs more consistent and reusable.
- Instrumented workflow runs with LangSmith for trace-level debugging and prompt iteration.
Python
LangGraph
Pydantic
LangSmith
LLM Evaluation
$ query inventory12 guarded tools4 warehouses
AI ApplicationGitHub
AI-Powered Inventory Management
Built a FastAPI inventory system with PostgreSQL, MongoDB audit logs, JWT auth, and a conversational agent for natural-language operations.
- Managed 22 products across 4 warehouses.
- Designed 12 permission-aware agent tools.
- Separated transactional data from flexible audit logging.
FastAPIPostgreSQLMongoDBLangChainOllama
2.3x
multi-node throughput gain
Data EngineeringGitHub
Distributed Data Processing & Scalability Analysis
Built Dockerized single-node and multi-node Spark environments and benchmarked PySpark ETL performance on NYC Taxi data.
- Measured runtime, CPU, memory, and throughput trade-offs.
- Built reproducible Spark experiments in Docker.
- Showed when cluster parallelism improves processing performance.
PySparkApache SparkDockerETLBenchmarking
Experience
Applied research, teaching, and ML delivery.
Mar 2026 — Present
AI R&D Intern · Biomedical Sensors & Systems Lab, University of Memphis
Building reproducible biomedical ML workflows for EEG signal processing, sleep spindle detection, and motor imagery classification.
- Audited 109 PhysioNet EEG subjects and established a clean 106-subject baseline.
- Applied LOSO validation with CSP + LDA and CSP + linear SVM for realistic multi-session benchmarking.
- Generated PSD visualizations and confusion matrices to support signal-quality analysis and technical reporting.
PythonMNESignal Processingscikit-learnResearch Documentation
Jan 2024 — Dec 2025
Data Science Fellow / Graduate Assistant · University of Memphis
Supported graduate-level coursework in applied statistics, machine learning, and biostatistical methods through hands-on coding templates and mentoring.
- Developed R starter code and assignment templates for Advanced Statistical Learning II.
- Mentored 30+ students in SAS-based statistical analysis and ML implementation.
- Translated modeling concepts into reusable, reproducible learning workflows.
RSASStatisticsMachine LearningMentoring
Jan 2023 — Dec 2023
Machine Learning Intern · Microline Information Systems
Worked on supervised ML modeling, data preprocessing, workflow automation, and early-stage REST integration for internal classification use cases.
- Built and evaluated supervised models with Python and scikit-learn.
- Automated data-cleaning and model-evaluation scripts for internal reporting workflows.
- Explored API-based delivery of model predictions for prototype usage.
Pythonscikit-learnFeature EngineeringREST APIs
Credentials
Education and certifications.
M.S. Data Science
University of Memphis · Jan 2024 — Dec 2025
GPA: 3.93 / 4.00Machine LearningAdvanced Statistical LearningDatabase SystemsData MiningArtificial Intelligence
B.Tech Electronics & Communication Engineering
JNTU Hyderabad · Jul 2020 — May 2023
Foundation in signals, DSP, AI, DBMS, and embedded systems.Signals & SystemsDigital Signal ProcessingPythonAIDBMS