Open to Entry-Level Intern & FTE roles from May 2026
Hey there!
I’m Madhumitha.
A graduate student in Analytics at UC Berkeley, building things at the intersection of data, AI and real-world impact.
At heart, a data science and ML enthusiast with a bias for impact hooked on one question: what happens when you point good models at problems that genuinely matter? So far: early disease detection, mood-aware music and a lot of messy datasets made useful.
PythonMachine LearningLLMs & GenAISQLExcelPower BIAPI DevelopmentStorytellingProduct MetricsKPI Tracking
4+
Projects
1
IEEE Paper
2
Internships
I’m Madhumitha, a graduate student in Analytics at UC Berkeley, with a B.Tech in Information Technology from SSN Chennai. I’ve been building things at the intersection of data, AI and real-world impact since before “GenAI” was a buzzword.
I’ve always been curious about systems: how they work, how they break and how they can be improved. That curiosity is what draws me to analytics and AI. I love working with data, uncovering patterns and turning insights into ideas that eventually become real products. What excites me most is using emerging AI technologies to build solutions that make their way into the real world.
I’m particularly fascinated by AI strategy: how organizations adopt new AI capabilities, redesign workflows and build entirely new products around them. That intersection of technology and business is something I constantly explore, and it shapes how I think about every project I take on.
Outside of work, I’ve cared deeply about the environment since I was around ten years old. That long-standing interest in community and environmental impact informs how I think about responsible innovation and the long-term consequences of what we build.
When I’m not working with data or exploring new AI tools, you’ll probably find me playing table tennis, a sport I once played competitively and still love for the focus, agility and strategy it demands. Or exploring cafes, trying new food or unwinding by the beach.
I’m a builder at heart. I love immersing myself in projects, experimenting with new tools and staying deeply engaged with what I’m creating. Looking ahead, I’m excited about contributing to meaningful systems and someday taking that further by founding a company that uses technology and data to create real-world impact.
Currently learning
How AI actually changes organizations — not the hype, the real redesign of workflows, decisions and products.
Happiest when
Something I built is being used by real people and making their lives a little easier.
Would drop everything for
A hard problem with real stakes. The messier the data and the higher the impact, the more interested I am.
Secret weapon
I taught myself an entirely new programming language in 2 months just by reading my internship codebase. I don’t wait to be taught.
Languages & Frameworks
Core Stack
PythonSQLJavaScriptHTML/CSSC++JavapandasNumPyscikit-learnTensorFlowmatplotlibseabornMySQLOracle DBSpring BootHibernate
AI & Machine Learning
What I Build With
LLMs & GenAITransformersSupervised LearningUnsupervised LearningNLPFine-TuningPrompt EngineeringOpenAI APIAnthropic / Claude APIHuggingFaceTime-Series AnalysisForecastingYOLOCollaborative FilteringPredictive ModelingStatistical AnalysisComputer VisionOpenCV
Cloud, Data & MLOps
Infrastructure & Delivery
DatabricksAzure SQLSupabase (PostgreSQL)REST APIsGitETL Pipelinesn8n (Workflow Automation)Model Training & EvaluationModel GovernanceData Quality & ValidationReal-Time API IntegrationData ModelingData WarehousingReproducible Analysis
Analytics & Visualization
Insight & Reporting
Power BIExcelExploratory Data AnalysisDashboard DevelopmentDAX & Data ModelingBusiness IntelligencePivotTablesPower QuerySUMIFS / VLOOKUP / XLOOKUPData ProcessingData Ingestion & TransformationData Quality Checks
Professional Skills
How I Work With People
Analytical RigorBusiness StorytellingProblem SolvingCross-functional CollaborationRequirements GatheringStakeholder ReportingProject ManagementInnovative ThinkingGrowth MindsetOwnership
AI-Driven Prognostic Monitoring and Personalized Intervention Platform for Alzheimer’s
CNN-LSTM · Transformers · Computer Vision · MRI Analysis · PyTorch
Overview
Deep learning based Alzheimer’s disease classification from MRI brain scans, identifying stages: Non-Demented, Very Mild, Mild and Moderate Dementia. Published in IEEE Xplore at the 2025 ICCDS.
My Role
Part of a team. My contributions focused on time-series analysis, LSTM modeling and building the clinician-facing application interface.
· Correlated CNN-extracted features with longitudinal OASIS dataset data to model disease evolution over multiple time points
· Prepared MRI feature vectors into patient-level time-series sequences capturing temporal patterns including cortical thinning and ventricular dilation
· Implemented LSTM networks to learn long-term sequential dependencies and avoid vanishing gradient problems
· Correlated MRI features with cognitive scores (MMSE, ACE) and clinical ratings to build prediction patterns between brain atrophy and cognitive impairment stages
· Built the model to forecast future atrophy patterns and cognitive decline trajectories, supporting early intervention and personalized treatment planning
· Developed the clinician-facing app: MRI upload and auto-classification (AD / CN / EMCI / LMCI), patient history tracking and confidence scores
· Enabled side-by-side comparison of MRI-based classification with longitudinal assessments for comprehensive diagnosis support
What It Does
· Classifies MRI scans into 4 Alzheimer’s progression stages using a CNN-LSTM hybrid
· Captures spatial patterns (CNN) and longitudinal progression signals across slices (LSTM)
· Performs image preprocessing and augmentation: normalization, resizing, dataset balancing
· Provides a visual analytics interface for clinicians highlighting affected brain regions
· Achieved 87% classification accuracy across dementia progression stages
CNN-LSTMMRI ClassificationImage AugmentationPyTorchVisual AnalyticsGit
Results
Moodify: A Spotify-Integrated Framework for Real-Time Emotion-Driven Music Recommendation
Python · CNN · PCA · K-Means · OpenCV · Haar Cascade · Spotify API · FER2013
Overview
Moodify is a real-time emotion-driven music recommendation system that uses facial expression recognition to detect the user’s current mood and automatically generate a Spotify playlist tailored to it.
My Role
Part of a team. My contribution focused on the emotion detection and classification model pipeline.
· Integrated the Haar Cascade Classifier to detect faces from live webcam video frames in real time
· Built the CNN-based emotion classification model trained on FER2013 (35,887 grayscale images) to classify 7 emotions: Angry, Disgust, Fear, Happy, Neutral, Sad and Surprise
· Implemented frame-level emotion tracking across 50 frames to determine the dominant emotion per session
· Handled image normalization, grayscale conversion and dimensional formatting for model inference
· Ensured the detected emotion label was correctly saved and passed downstream to the playlist mapping pipeline
Pipeline
1
Haar Cascade Classifier detects faces from live webcam frames
2
CNN classifies emotion into 7 states with confidence score
3
PCA reduces audio feature vectors to 6 components for efficient clustering
4
K-Means groups Spotify tracks into 7 mood-based clusters using audio features
5
Detected emotion maps to mood cluster and 26-song playlist is auto-created via Spotify API
Emotion → Playlist
😊 Happy → Upbeat
😢 Sad → Calming
😠 Angry → High-energy
😱 Disgust → Soothing
😲 Surprise → Dynamic
😓 Fear → Ambient
😐 Neutral → Balanced
PythonCNNFER2013OpenCVHaar CascadePCAK-MeansSpotify APIFlask
GitHub ↗
Results
Autonomous Chess Piece Navigation System
Python · OpenCV · Stockfish AI · FEN Notation · Microcontrollers · Robotics
Overview
AI-driven chess system capable of detecting board states, computing optimal moves and executing them physically via a robotic arm in real time. Funded by SSN Research Centre.
My Role
· Built the computer vision pipeline that translates the physical board state into a digital representation using OpenCV: chessboard grid detection, square segmentation and piece recognition in real time
· Converted detected board configurations into FEN notation for integration with the Stockfish chess engine to compute optimal moves
· Contributed to designing the pipeline that maps AI-generated moves to board coordinates, enabling the system to translate strategic decisions into physical piece navigation
· Demonstrated how computer vision, AI decision systems and physical game environments can be integrated to create interactive intelligent systems
What It Does
· OpenCV detects chessboard grid, identifies piece positions and converts board state to FEN notation
· Stockfish engine computes optimal moves from the detected board configuration
· State-to-action pipeline maps digital moves to physical board coordinates for robotic navigation
· Real-time board recognition updates game states dynamically as moves are played
· Demonstrates full integration of computer vision, AI decision systems and physical automation
PythonOpenCVStockfish AIFEN NotationMicrocontrollersRoboticsComputer Vision
GitHub ↗
Results
QueryMind — AI-Powered Natural Language Data Assistant
Claude AI · n8n · Supabase · PostgreSQL · Python · REST APIs
Overview
An end-to-end AI-powered analytics tool that converts plain English business questions into live SQL queries and actionable insights. Built out of a personal interest in applying emerging AI technologies to real-world data workflows.
What It Does
1
Accepts a natural language business question via a branded dashboard
2
Claude AI API generates optimized PostgreSQL from the question automatically
3
Query executes live against Supabase PostgreSQL (Olist dataset — 100K+ orders, 9 relational tables)
4
Returns structured analysis: plain-English summary, data quality audit and visualization recommendation
5
n8n orchestrates the full workflow end-to-end with no manual intervention
A/B Prompt Engineering Study
Conducted a rigorous A/B test comparing a minimal system prompt against a schema-aware engineered prompt across 20 diverse business questions — measuring SQL accuracy, query correctness and output reliability. The engineered prompt achieved significantly higher accuracy, providing evidence-based justification for the final system design.
Key Outcomes
· End-to-end NL-to-SQL pipeline with live database execution and structured output
· Schema-aware prompt engineering validated through experimental A/B design
· Demonstrates proficiency in AI tool integration, automated pipeline development and translating technical outputs into business-ready insights
· Custom-built branded dashboard for non-technical business users
Claude AI APIn8nSupabasePostgreSQLPrompt EngineeringREST APIsA/B TestingNL-to-SQLWorkflow Automation
GitHub ↗
Results
Jun – Aug 2024
FinTech
FinTech
Technology Summer Intern
Barclays · Private Bank Processing Team · Pune, India
Worked on the Document Management & Workflow (DMW) platform used by relationship managers to manage high-net-worth client requests account maintenance, fund transfers and documentation workflows.
· Designed and implemented Oracle SQL views to consolidate fragmented customer and account data across multiple tables.
· Integrated views into the Hibernate-based backend, modifying POJOs and data access layers for view-based data consumption.
· Refactored data loading to minimize repeated database calls, loading datasets once into memory at server init, improving performance.
· Transitioned manual stub sheet config to a database-driven implementation, enabling dynamic retrieval and better pipeline maintainability.
· Collaborated with developers and QA to validate queries, backend integration and data consistency across testing environments.
· Contributed to digital automation initiatives reducing manual paperwork for private banking operations.
· Implemented SonarQube static code analysis, identifying 20+ code quality issues including maintainability risks and duplicated code.
· Improved code maintainability metrics and reduced technical debt by resolving critical and major SonarQube issues across backend modules.
Oracle SQLJavaHibernate ORMJiraGitPrivate Banking
Certificate of Completion
Oct – Nov 2023
Public Sector
Public Sector
Data Analytics Intern
Chennai Metro Rail Limited · Revenue Department · AFC System · Chennai, India
Analyzed metro ridership and ticketing data to identify patterns in passenger demand, revenue generation and operational efficiency across routes.
· Performed data analysis on ticketing datasets to study passenger demand, peak-hour patterns and station-level traffic.
· Built Excel-based dashboards using PivotTables and advanced formulas to visualize ridership and revenue trends.
· Automated daily/weekly reporting workflows, reducing manual reporting effort through structured Excel templates.
· Prepared visual summaries for internal stakeholders, making technical insights accessible to non-technical decision-makers.
ExcelPivotTablesVLOOKUPTrend AnalysisKPI TrackingBI Reporting
Certificate of Completion
Master of Analytics
University of California, Berkeley
Fall 2025 Coursework
Python for AnalyticsOptimization AnalyticsRisk Modeling, Simulation & Data AnalyticsML & Data AnalyticsAnalysis & Design of Databases
Spring 2026 Coursework
Economics of Supply ChainMachine Learning & Data Analytics 2Introduction to Data Modeling, Statistics & System SimulationTransportation Analytics
B.Tech Information Technology
Sri Sivasubramaniya Nadar College of Engineering
Relevant Coursework
Programming & Design PatternsDatabase TechnologyData Communication & NetworksAutomata TheoryProbability & StatisticsData Analytics & VisualizationOperating SystemsArtificial IntelligenceInternet of ThingsWeb ProgrammingCloud & Distributed ComputingDeep LearningComputer VisionMachine LearningFull Stack Development
Certifications
Postman API Fundamentals Student Expert
Postman · API Development
Introduction to Internet of Things
CISCO Networking Academy
Fuzzy Sets, Logic & Systems 90%
NPTEL · IIT Coursework
Microsoft Power BI Data Analyst
Coursera · Microsoft · In Progress