SYSTEMS NOMINAL / INDIANAPOLIS, IN 39.77°N 86.16°W / --:--:-- LOCAL

Roshan Naidu

Data Scientist · AI/ML Engineer · Researcher

Now
MS Applied Data ScienceIndiana University Indianapolis · GPA 4.0/4.0 · May 2027
Role
Data Engineer InternIU School of Medicine · plus Graduate Research Assistant
Research
3 IEEE papersICALT 2026 · ICOSEC 2024 · INCET 2023
Awards
Hackathon winner1st place + Best UI/UX, IU Indy CS Club · Hack-Earth finalist
SCROLL TO FLY

[01] Mission log 06 systems

Experience & leadership

Six star systems, one for each role. Keep scrolling to jump between them at light speed.

  1. ActiveSep 2026 — Present

    Data Engineer Intern

    Indiana University School of Medicine · Indianapolis, IN

    • Engineering a SMART-on-FHIR application on Epic and FHIR R4 to automate MASLD risk screening.
    • Developing FIB-4 risk scoring and rule-based triage logic for automated patient risk classification and referral alerts.
    • Integrating Epic EHR data for real-time processing, risk assessment, and clinical decision support.
    • Running pilot tests and folding user feedback into the workflow; preparing a research manuscript.
  2. ActiveSep 2026 — Present

    Funding Board Member

    Indiana University Indianapolis · Indianapolis, IN

    • Reviewing and evaluating funding requests from student organizations and campus initiatives through the SOG (Student Organization Grant) and GPEG (Graduate and Professional Education Grant) programs.
    • Assessing proposals for alignment with student needs and campus priorities.
    • Working with fellow student representatives and board members to reach collaborative funding decisions.
    • Building experience in organizational governance, budgeting, and financial decision making.
  3. ActiveSep 2025 — Present

    Graduate Research Assistant

    Indiana University · Indianapolis, IN

    • Deployed a deep neural network web application that detects and classifies obstructions in rivers and streams across Indiana.
    • Developed an Empathy AI game combining AR/VR with AI/ML to study human–machine empathy.
    • Built quantum machine-learning models that integrate quantum principles with AI/ML.
    • Collected and processed real-time research data and co-authored research articles with faculty and students.
  4. Jan 2026 — Aug 2026

    President

    Data Science & Machine Learning Club at IU · Indianapolis, IN

    • Led the executive board and general members in planning and delivering club events.
    • Ran weekly meetings and coordinated students, faculty, and industry executives.
    • Managed event logistics, scheduling, and operations for campus programming.
  5. Jan 2025 — Jun 2025

    Data Analyst Intern

    Masycoda Solutions Pvt. Ltd. · Nagpur, India

    • Built interactive Power BI dashboards to support business decisions.
    • Analysed large datasets with DAX and Power Query, and normalised high-volume sales tables to cut redundancy.
    • Improved dashboard efficiency by up to 50% with filters, slicers, and drill-through.
  6. Jan 2023 — Apr 2024

    Technical Head

    Darlins Production · Chennai, India

    • Designed and launched the website with SEO optimisation to grow organic traffic.
    • Resolved 95% of customer queries within 24 hours.
    • Developed a brand strategy that unified the company’s online presence.
Skip the flight ↓

[02] Mission archive 07 projects

Selected work

Seven builds across deep learning, geospatial AI, real-time systems, and data engineering. Five are live — open them.

Research · Geospatial AILive

Indiana Stream Blockade Detection

Research application from my graduate assistantship under Dr. Sunandan Chakraborty. It runs a trained PyTorch model over georeferenced GeoTIFF imagery to flag candidate obstructions in Indiana rivers and streams, then maps detections back onto official NHD flowlines.

  • 9-channel input stack: RGB, NIR, NDVI, water proxy, DEM, slope, brightness
  • Decision gate that blocks bridges, riffles, canopy shadow and other false positives
  • Interactive map and GeoJSON export, with human review required before any record is official
  • PyTorch
  • Streamlit
  • GeoTIFF
  • NHD hydrography

Winner · Best UI/UXLive

CodeSafe

A whiteboard for developers: a browser-based workspace where a team edits the same project in real time, Google Docs style. First place and the Best UI/UX award at the Computer Science Club Hackathon at IU Indianapolis, 2026.

  • Conflict-free concurrent editing with Yjs CRDTs bound to the Monaco editor
  • Shared file tree, live chat, suggestion voting, and a collaboration leaderboard
  • Viewer / editor / admin roles, MongoDB persistence, in-browser JS and Python console
  • React
  • TypeScript
  • Yjs
  • Node.js
  • WebSockets
  • MongoDB Atlas

Finalist · SoloLive

Hack-Earth: Energy Efficiency Model Lab

An interactive ML lab for modelling building heating and cooling loads, built solo for the Hack-Earth hackathon by Allstate, where it reached the finals. It trains and compares models side by side while tracking the carbon cost of the compute itself.

  • Regression and classification across linear models, random forests, XGBoost, LightGBM and CatBoost
  • CodeCarbon emissions tracking for training and inference
  • What-if scenario tool for testing building designs
  • Streamlit
  • scikit-learn
  • XGBoost
  • LightGBM
  • CatBoost
  • CodeCarbon

Workshop · Quantum MLPresented

Quantum Machine Learning Workshop

A hands-on workshop I designed and presented at the Data Science & Machine Learning Club event at Indiana University. Attendees take a circuit from OpenQASM into PennyLane and watch a quantum verdict drive a classical decision.

  • Small OpenQASM 2 parser that rebuilds the circuit gate by gate in PennyLane
  • 4-qubit demo that classifies a hidden function as constant or balanced
  • Worked example that routes audit cases based on the circuit’s result
  • PennyLane
  • Qiskit
  • OpenQASM
  • Jupyter

Data Science · INFO-H 501Live

Music & Mental Health Dashboard

A multipage analytics app exploring how listening habits relate to self-reported anxiety, depression, insomnia and OCD, built on a public Kaggle survey. Designed with clear caveats: observational, not diagnostic.

  • Interactive exploration by genre, listening hours and demographics
  • Mood analyser page and statistical modelling with statsmodels
  • Explicit treatment of sampling bias and sensitive self-reports
  • Streamlit
  • pandas
  • Plotly
  • statsmodels

Database Design · LIS-S 511Live

Urban Wildlife Rescue Management System

A relational database designed for wildlife rescue operations, from incident report through treatment to release, with a read-only demo app on top. Reporter, responder and staff contact details are deliberately hidden from public tables.

  • 15-table normalised MySQL schema covering incidents, animals, facilities, permits and inventory
  • Searchable registers with filtered CSV export
  • Analytics by incident type, species, health condition and status
  • MySQL
  • SQLite
  • SQL
  • Streamlit

[03] Signal log 03 publications

Peer-reviewed research

Three IEEE conference papers across two universities. ORCID 0009-0009-5595-4964 ↗

  1. 2026IU

    Exploring Educators’ Perspectives on VR and AI for Social-Emotional Learning in Middle Schools

    Shreya Kamath, Roshan Naidu, Shashi Kant Shankar, Harley Locklear, Karuna Saravanan, Ekaterina Muravevskaia

    2026 IEEE International Conference on Advanced Learning Technologies (ICALT) · pp. 375–376 · Indiana University

    DOI ↗
  2. 2024VIT

    Deep Convolutional Neural Network based Wheat Leaf Disease Detection in Precision Agriculture

    Sreyashi Dey, Roshan Naidu, Om Subrato Dey, A Balasundaram, Ayesha Shaik

    2024 5th International Conference on Smart Electronics and Communication (ICOSEC) · pp. 1183–1188 · VIT

    87.31% accuracy classifying healthy wheat against three diseases across 4,800 images. Code ↗

    DOI ↗
  3. 2023VIT

    Advanced Traffic Sign and Lane Discernment

    Roshan Naidu, Vedhapriyavadhana R

    2023 4th International Conference for Emerging Technology (INCET) · pp. 1–5 · VIT

    99.69% accuracy in a CNN computer-vision model for ADAS under adverse conditions. First author.

    DOI ↗

[04] Onboard systems 04 modules

Skills & education

> languages

  • Python
  • R
  • SQL
  • C++
  • HTML/CSS
  • Bash

> data_science_stack

  • PyTorch
  • TensorFlow
  • scikit-learn
  • Pandas
  • NumPy
  • OpenCV
  • NLTK
  • Plotly
  • Matplotlib
  • Seaborn
  • Streamlit

> tools

  • Power BI
  • Tableau
  • Excel
  • Git
  • Roboflow
  • Quadratic AI
  • LLMs

> platforms

  • Docker
  • Hugging Face
  • Postman
  • Qiskit
  • VS Code
  • StarUML
  • Wolfram

Aug 2025 — May 2027

Master of Science, Applied Data Science

Indiana University Indianapolis

4.0/4.0 GPA

Jul 2021 — Aug 2025

B.Tech, Computer Science with AI & Machine Learning

Vellore Institute of Technology, Tamil Nadu

8.7/10 CGPA

Honors

  • Winner + Best UI/UX Computer Science Club Hackathon at IU Indianapolis, 2026 — CodeSafe
  • Finalist Hack-Earth 2026 — reached the finals as a solo developer against hundreds of teams and participants

[05] Open a channel channel open

Let’s build something that matters.

I am finishing my MS in May 2027 and always glad to talk about data science, machine learning, and research. Email is the fastest way to reach me.