Named Entity Recognition for Vehicle Attributes Extraction
NLP pipeline for extracting vehicle attributes from text descriptions
Built an NLP pipeline on the FindVehicle dataset with 42K descriptions and 21 entity classes. Converted CoNLL-style annotations to BIOE tagging and benchmarked RNN, LSTM, BiLSTM, and transformer architectures. Fine-tuned RoBERTa-base with AdamW, outperforming all baselines.
B
Big Data-Driven Health Risk Assessment
Wearable health data processing and risk segmentation using Spark
Processed 141M+ wearable health records totaling 17GB using PySpark on HDFS with Parquet format. Engineered risk features and identified key correlations across 127K+ users. Built K-Means segmentation and Linear Regression risk-scoring model deployed via Streamlit for real-time prediction.
C
Content and Behavior-Aware Movie Recommendation System
- Designed a full-stack web app for customized movie recommendations, leveraging subtitle data from IMDb API.
- Integrated both content-based and behavior-aware filtering to deliver tailored suggestions.
- Built an NLP pipeline to extract genre and summary metadata from subtitles.
- Analyzed implicit user behavior to support session-aware and dynamic content personalization
- Deployed the application on Amazon Amplify, enabling fast, real-time recommendations with 80% accuracy and maintaining computational efficiency.
J
Job Scam Detection in Gmail
- Built ML system to detect fraudulent job emails using Enron, survey, and synthetic data
- Engineered 15+ features (e.g., scam/urgency keywords, sender metadata) & trained models (LogReg, SVM, RF, XGBoost)
- Achieved 96.7% accuracy & 73.4% F1 with Random Forest; used SHAP for interpretability
- Applied SMOTE + Stratified K-Fold CV to address class imbalance
- Selected as a ChatGPT Ambassador through the CSU × OpenAI partnership, supporting the rollout of ChatGPT Edu to 460,000+ students across 23 CSU campuses.
- Planned and conducted “Campus Pulse,” an interactive on-campus Q&A session with students focused on real-world AI usage, challenges, and expectations.
- Collected actionable feedback and student reviews on AI performance, usability, and academic impact to inform responsible AI adoption.
- Educated students on ethical, healthy, and safe AI use within an academic environment.
Evaluated over 100 graduate-level assignments on ML algorithms, Python, and SQL for technical correctness and code quality. Provided feedback on feature engineering and model selection. Reviewed database design including ER/EER models, normalization, and advanced SQL concepts like window functions.
Designed a hybrid RAG architecture combining Pinecone vector search and Neo4j knowledge graphs with evaluation metrics for scalable chatbot responses. Collaborated with the RLHF team to align model outputs with human feedback and created workflow visualizations that reduced onboarding time by 30%.
Conducted keyword analysis, competitor benchmarking, and audience segmentation to support content strategy. Implemented on-page and off-page optimization strategies, improving mobile page speed from 15 to 65 and desktop from 25 to 75. Reduced page load time from 4 seconds to 1 second and cut bounce rate by 60%.
Data Analyst Trainee (SEO)
Dec 2018 — Jan 2019 · Chennai, Tamil Nadu, India
• Worked closely with the development team, and helped extract all the urls from the main website using python
• Learned to use Google Analytics tools, and ways of data extraction from it
• Learned client server relationship and http client
• Learned to use various SEO tools, plugins and data extraction methods
Specialization Certification in Advanced Digital Marketing and Growth Strategies
Nov 2023
Advanced Digital Marketing course offered by Wharton Online, helped me gain a profound understanding of digital marketing strategies and tools. This comprehensive program covered strategic planning, content marketing, SEO, social media marketing, email marketing, PPC advertising, analytics, and conversion rate optimization. I also learned to stay up-to-date with the latest industry trends. Real-world case studies provided practical insights, ensuring a well-rounded education. Overall, the course has equipped me with the skills and knowledge to excel in the dynamic field of digital marketing, from crafting effective strategies to implementing data-driven decision-making.
Bachelor of Technology - BTech in Information Technology
Grade: 3.4
Activities and societies: Smart India Hackathon, Comparing for various college functions , Student president of information Technology Department, Industrial Visit coordinator, Cultural club leader.
Credentials
N
Course
Fundamentals Of Deep Learning
NVIDIA Deep Learning Institute
Issued Jan 2026
M
Course
Perform analytics in Power BI
Microsoft
Issued Oct 2024
U
Course
Python Specilization - Capstone project
University of Michigan
Issued Jul 2024
U
Course
Using Databases with Python
University of Michigan
Issued Jul 2024
Recommendations
No recommendations yet
Nothing posted yet.
Skills
EnglishHadoopReact.jsAmazon Web Services (AWS)Node.jsHTML & CSSBERT (Language Model)C++Apache SupersetMicrosoft Power BIPandas (Software)TableaudockerApache AirflowGoogle Cloud Platform (GCP)Core Java