I'm a software engineer with 3+ years of experience designing, building, testing and supporting software, from backend services and data platforms to machine-learning and AI applications.
Today I build and maintain backend and data-processing services for an enterprise healthcare platform at Insight Global, working in Python, SQL and PySpark on Databricks, Snowflake and AWS. Most of my work is writing reusable, well-tested components, tracing issues across distributed systems to the root cause, tuning performance, and shipping changes through code review and CI/CD.
- 🧰 Daily tools: Python · Java · SQL · PySpark · Databricks · Snowflake · AWS · Docker · Git
- 🤖 Lately: incremental Snowflake pipelines (Snowpipe, Streams & Tasks, SCD Type 2, reconciliation), clinical reporting on EHR data models (SQL, Oracle, CCL), AI agents with guardrails and evaluation (Claude API, MCP), and forecasting and optimization
- 📄 IEEE-published co-author (NLP, 93.82% accuracy)
- 🎓 M.S. Computer Science, NJIT · B.Tech Computer Science and Engineering, Amrita Vishwa Vidyapeetham
- 🔎 Open to software engineer, backend and full-stack roles
| Project | What it does | Tech |
|---|---|---|
| ❄️ Snowflake Incremental Data Warehouse | Daily broker account and trade files loaded by Snowpipe (REST API), Streams and a scheduled 5-task graph into an SCD Type 2 account dimension, a trade fact and a daily position snapshot. Every file is reconciled to the cent against the sender's control totals (60/60 files, $7.3B notional), all 178 planted bad rows are caught with a reason, and an Oracle PL/SQL rebuild matches count for count. 62 tests. Synthetic data. | Snowflake · Snowpipe · Streams & Tasks · Snowpark · SQL · Oracle PL/SQL · Python |
| 📦 Distribution Center Demand Forecast & Capacity Optimization · dashboard | Forecasts 12 weeks of regional demand (about 5.8% error vs 9.7% for a seasonal baseline), then a linear program assigns volume across 4 distribution centers within capacity at minimum cost. Compares capacity-expansion scenarios. Network inputs are documented assumptions. | Python · Prophet · ARIMA · PuLP · SQL · Tableau |
| 🏥 Clinical Operations Reporting (Cerner Millennium-style) · case study | 7 KPI and operational reports (ED boarding, lab turnaround, critical-result calls, readmissions, census) on a 13-table Millennium-style data model, in SQL and CCL with HTML/MPage output. 21 pre-publish checks, 19 tests, Oracle port reconciled 238/238 values, and a Discern rule tested in silent mode (60-min window: 71 false alerts vs 993). | SQL · Oracle · CCL · Python · HTML |
| 🤖 AI Data Analyst Agent | Claude agents turn plain-English questions into validated SQL, charts and reports through a read-only MCP server. CI, tests and a 30-question evaluation (28/30 correct). | Python · Claude API · MCP · PostgreSQL · Streamlit |
| 📞 Call Center KPI Analytics · dashboard | Workforce-management KPIs (AHT, service level, abandonment, occupancy, shrinkage) at 30-minute intervals for 3 queues, with Erlang C staffing. Found a late-morning service-level dip with flat staffing, pointing to schedule shifts rather than more headcount. | SQL · Python · Tableau |
| 🐝 HiveQL Supply-Chain Analytics | Retail demand and forecast backtests in HiveQL on Apache Hive 4: ORC tables partitioned by region and year, a bucketed table, window functions and CTAS. Results match the original Python pipeline. | HiveQL · Apache Hive · Docker |
| 🛡️ Credit-Card Fraud Detection | Benchmarks Logistic Regression, Random Forest and Gradient Boosting on imbalanced data; best model served by a FastAPI scoring service. | Python · scikit-learn · FastAPI |
| 📚 RAG Document Assistant | Retrieval-augmented assistant with embeddings, cosine-similarity retrieval and grounded answers with citations. | Python · Embeddings · FastAPI |
| 📝 SecureComment (IEEE) | TF-IDF with voting and stacking ensembles for toxic-comment classification. Basis of a co-authored IEEE paper. | Python · NLP · scikit-learn |
| ☁️ AWS Image Recognition Pipeline | Event-driven pipeline: SQS decouples parallel workers running Rekognition object detection and writing results to S3. | AWS · boto3 · SQS |
| 🗄️ Hadoop MapReduce | Distributed batch processing on Hadoop/HDFS with a combiner for local pre-aggregation; deployable on AWS EMR. | Java · Hadoop · AWS |
| 📈 Data Mining & Predictive Analytics | Benchmarks Random Forest, SVM, Gradient Boosting and a neural net with cross-validation and automatic best-model selection. | Python · scikit-learn |
| 🚗 Rent-A-Car Database System | 3NF schema, analytical SQL and a Streamlit reporting dashboard. | SQL · Python · Streamlit |
| 📊 Web Scraping & Analytics in R | Paginated scraping with rvest, tidy cleaning and ggplot2 visualizations. | R · rvest · ggplot2 |
- Software Engineer, Insight Global (May 2025 to present): backend and data-processing services for an enterprise healthcare platform (Python, SQL, PySpark, Databricks, Snowflake, AWS)
- Software Engineer Intern, Genpact (2024): Python data-processing and ML components, model comparison and testing
- Software Engineer, Research Assistant, Amrita Vishwa Vidyapeetham (2023 to 2024): built the NLP pipeline behind the IEEE-published SecureComment system
- Software Engineering Specialist, Chegg (2023): reviewed and debugged 300+ Python, Java and SQL solutions
- Software Engineer Intern, Genpact (2022): Python and SQL components and scheduled Azure Data Factory pipelines
Languages: Python · Java · SQL · Oracle SQL · PL/SQL · HiveQL · CCL (Cerner) · JavaScript · R · C++
Backend & engineering practices: REST APIs · Microservices · OAuth 2.0 / JWT · RBAC · Docker · Git · Code review · Unit & integration testing · CI/CD (GitHub Actions) · Agile
Cloud & data: AWS · Azure · Oracle Cloud (OCI) · Oracle Database · Databricks · Snowflake · Snowpipe · Streams & Tasks · Snowpark · PySpark · Apache Spark · Hadoop · Apache Hive · PostgreSQL · MySQL · SQLite
AI & ML: scikit-learn · NLP · LLMs · Claude API · AI agents · MCP · RAG · Time-series forecasting · Linear programming
Analytics & BI: Tableau · Power BI · Pandas · Matplotlib · ggplot2 · KPI & operational reporting · Data validation
Healthcare data: Cerner Millennium data model · Discern Rules · MPages · Clinical KPIs
- Oracle: 7 certifications across OCI, AI, data platform, ERP and AI Database
- Databricks: Advanced Machine Learning Operations · Fundamentals · Fine-tuning Embeddings and Advanced Retrieval
- Google Cloud: Generative AI · Large Language Models · Prompt Design in Vertex AI
- Microsoft Learn: 31 learning paths and 109+ badges across Azure, security, data and AI
- Anthropic: 14 Claude courses (AI Fluency, Claude API, MCP, agents, Claude Code)
- HackerRank: SQL (Advanced) · Software Engineer · Software Engineer Intern · Frontend Developer (React)
- LeetCode: 100 Days Badge 2026
- Salesforce Trailhead and Forage job simulations (BCG, Deloitte, Quantium and others)
📰 SecureComment: Safeguarding Online Discussions with Intelligent Toxic Comment Filtering, IEEE IATMSI 2024 (co-author)
