Mechatronics engineer with a combined interest in Machine Learning/AI, Data Science, and Robotics.
Built end-to-end ML & DL projects spanning SQL analytics, Power BI, Streamlit dashboards, classification pipelines,
NLP sentiment analysis, and an LLM-powered policy analyzer. Engineering experience includes a PID-stabilized quadcopter and
ROS/Gazebo work with a robotic arm, UAV and mobile robots.
Open to opportunities spanning Robotics Research, AI/ML, and Data Science.
Relevant coursework: International Marketing, Industrial & Logistics Management, Strategic Management, Digital Business Models, Risk Management in Global Supply Chains, Corporate Governance.
Active participation in case studies, business simulations, and cross-cultural team projects.
Master's Thesis: Enabling a Sustainable Circular Economy through Industry 4.0 : Innovations, Challenges, and Future Prospects.
Project works:
Study of supply chain risk management in Toyota, Apple, and Unilever.
Study of the Evolution of Marketing.
Diversity in Human Resource Management.
Bachelor of Technology in Mechatronics Engineering with specialization in Robotics
Jun 2018 – May 2022 Chennai, India
SRM Institute of Science and Technology
Relevant coursework: Robotics & Mechatronics, Embedded Systems, Control Systems, Machine Design, Power Electronics.
Final-year project: Design and Fabrication of Smart Helmet Designing – focused on rider safety and embedded electronics.
Built a strong foundation in mechanical design, embedded programming, and automation systems — spanning both coursework and self-directed projects.
Projects
Police Traffic Stop Analytics (SecureCheck)
Streamlit dashboard on a MySQL traffic-stop database: every KPI, chart and table is driven by live SQL queries, with interactive Plotly charts and a vehicle look-up.
Wrote 20 SQL queries (aggregations, window functions, CTEs, joins) on an indexed schema — top vehicles in DUI stops, arrest rate by age group, high-risk violation patterns.
Analysed 16,146 district-level ICRISAT records (20 states, 311 districts, 1966–2017) on crop area, production and yield: Python cleaning, 15 EDA analyses and 10 business questions answered in MySQL.
Interactive Power BI report plus a Streamlit dashboard with year-range and state filters; yield is computed as total production ÷ total area, not as an average of yields.
Compared four classifiers on the IBM HR dataset (1,470 employees, ~16% attrition); selected Logistic Regression (AUC 0.82, recall 0.72) because high-accuracy models missed most leavers (Random Forest: 85% accuracy, 8.5% recall).
Feature engineering, preprocessing and model packaged in one scikit-learn pipeline behind a Streamlit app that scores attrition risk from 29 employee inputs.
Identified and removed a data-leakage feature that was inflating a second target to a near-perfect score; after the fix the target showed no real signal, so no prediction is offered.
Built and compared classification models (Logistic Regression, Random Forest, XGBoost) to predict Parkinson's, chronic kidney, and liver disease from clinical data.
Combined three independent, disease-specific pipelines into a single Streamlit application.
NLP pipeline (NLTK preprocessing, TF-IDF) classifying app reviews as positive, neutral, or negative, comparing classical ML (Naive Bayes, Logistic Regression, SVM) against a TensorFlow/Keras LSTM.
Used a leakage-safe train/test split at the unique-text level and documented the dataset's real limitations rather than overstating results.
Public Policy Insight & Impact Analyzer (AI Policy Analyzer)
NLP and LLM web app that turns long government bills (PDF, web link or pasted text) into plain-language summaries, industry impact reports, timelines, and short, medium and long-term impact assessments.
Combines rule-based legal NLP (section splitting, clause-type classification, Act and date extraction), TF-IDF/NMF topic modelling and Google Gemini analysis; LLM output is validated against Pydantic schemas.
Builds knowledge and section-dependency graphs (NetworkX, Graphviz) and a downloadable Markdown report in a Streamlit dashboard.