DarshanSenthil

Built a governed AI analytics product that translates business questions into validated SQL. Enforces read-only access, allowed-schema guardrails, and SQL validation before execution. Includes JWT authentication, data masking for sensitive fields, and transparent evidence so users can see exactly what SQL was run and why.

Built a real-time BTC-USD data pipeline that streams live price ticks from Coinbase, processes them through Kafka and Spark, transforms the data with dbt, and orchestrates everything with Airflow on a 5-minute schedule. The final output is a live Grafana dashboard showing price trends, volatility, bid/ask spread, and daily market summaries. Fully containerized with Docker.

Personalized fitness and nutrition tracking application that helps users plan workouts, analyze meals, and monitor progress based on individual goals, age, and activity level using AI-powered recommendations.

Hybrid recommendation system that combines content-based similarity with popularity signals to help users discover relevant books and courses, effectively handling cold-start scenarios with 1,200+ items across 8 categories.

AI-powered educational assistant designed to help users understand cancer-related topics with built-in safety guardrails. Features real-time chat, domain-constrained responses, and document processing capabilities.

Developed a modern data engineering pipeline to analyze Uber data using Google Cloud Platform, Python, Mage.ai, BigQuery, and Looker Studio for scalable data processing and real-time insights.

Performed web scraping for crime-related news headlines and post dates in New York, applied sentiment analysis for scoring, and generated a time-series plot to illustrate average sentiment scores over dates, offering insights into evolving news sentiments.

This movie recommendation system, built using KNN, Slope One, and SVD algorithms, offers personalized and accurate movie recommendations to users. Developed in Python, it uses large datasets and various metrics for evaluation. The project includes a web interface for users and documentation for easy setup and contributions.

The Online Auction website is developed using JavaScript, Java, SQL Workbench, and Tomcat Server. It offers a user-friendly interface with features like bidding, item listing, and user authentication. The project includes a documented database schema and scripts, enabling easy setup and contributions by developers.

Developed an informative dashboard to view the overall number of persons affected, recovered, and died by the devastating COVID-19 virus in a Demographic and Gender Level, as well as perform time-series analysis on this data.

The main objective of this research is to forecast the precision of a linear regression model based on the used-car sale price. Both buyers and sellers may utilize this model to make well-informed decisions about the price of a vehicle.

Built classifiers for identifying faces and categorizing numbers. The three classifiers were that we used are Naive Bayes, Perceptron, and K Nearest Neighbor.

Maze generation is done using DFS. The BFS and A star algorithms are used to identify a path for the agent to take in order to achieve the objective state.

Comprehensive dashboard tracking hiring efficiency metrics including average fill time, hires by month, and days to fill by recruiter/department for optimized recruitment processes.

Email marketing performance dashboard tracking key metrics like delivery rate, open rate, click-to-open ratio, and clickthrough rates across different campaigns.

Executive summary dashboard analyzing website traffic sources, session metrics, and performance across different channels including organic search and referral traffic.

Product line deep-dive dashboard with detailed sales analysis, geographic mapping, and category performance breakdown for comprehensive retail insights.

Executive summary dashboard providing comprehensive retail sales overview with total revenue metrics, geographic distribution, and overall performance indicators.

Interactive visualization of crime patterns and trends in New York City with demographic analysis and geographic mapping.

Comprehensive analysis of global climate patterns and temperature changes over time with interactive filters and trend analysis.

Real-time COVID-19 data visualization with demographic breakdowns, time series analysis, and geographic impact mapping.

Comprehensive analysis of NYC and San Diego accommodation patterns and booking trends with seasonal insights and pricing analysis.

Worldwide COVID-19 impact analysis with country-level data, infection rates, and comprehensive geographic visualization.