I build systems that solve real problems — production pipelines, automated outreach, and data-driven engineering at scale.
I am currently pursuing a MS in Computer Science at
Northeastern University.
Previously, I worked at
Amdocs Corporation for 1.5+ years where I built production
microservices for a telecom platform serving 10M+ subscribers with 99.5% uptime.
Experience
Where I've Worked
Amdocs Corporation — Pune, India
- Engineered 8+ production microservices using Spring Boot and Oracle for JCOM's CRM and telecom billing platform serving 10M+ subscribers, delivering 25+ features and maintaining 99.5% uptime across distributed service architecture
- Optimized Oracle database queries, achieving 60% performance improvement through composite indexing and query refactoring, eliminating 15 hours/week of processing delays and enhancing system responsiveness for critical operations
- Built automated test suites with JUnit and Mockito covering unit, integration, and API contract testing, achieving 80% code coverage and reducing production defects by 30%
- Developed 5 SOAP-based web services integrating Amdocs CRM with external BSS/OSS systems, processing 50K+ daily transactions with 99.7% uptime
- Containerized microservices using Docker and deployed on AWS EC2, reducing deployment time from 45 minutes to 8 minutes per release cycle
ISKCON Organization — Pune, India
- Automated infrastructure monitoring with Bash scripts and cron jobs, implementing alerting systems for CPU, memory, disk usage, and service availability that saved 4 hours weekly while reducing unplanned downtime by 25%
- Deployed cloud-native applications on AWS EC2 with ECR/ECS, reducing deployment time by 78% and achieving 93% uptime
- Designed normalized MySQL database schema (3NF) managing 100K+ records across devotee profiles, events, and financial transactions
Simba Developers Organization — Pune, India
- Architected distributed microservices using Node.js and Express for authentication and payment modules, enabling independent scaling and reducing cascade failures by 65% through fault-tolerant design patterns
- Integrated WebSocket modules for real-time bidirectional communication, achieving sub-50ms latency for live chat and push notifications across 10K+ concurrent connections
- Reduced app load time by 44% (3.2s → 1.8s) by optimizing Flutter state management with Riverpod and Bloc, implementing lazy loading, caching strategies, and widget tree restructuring
Projects
Things I've Built
Automated Job Hunt Pipeline
Production-grade data pipeline processing 8,000+ weekly postings across 6 sources. Star-schema analytics, TF-IDF fuzzy dedup, anomaly detection with SPC bounds, crash-safe WAL, circuit breakers, 8-layer email discovery, and pluggable validation pipeline.
Thyroid Disease Classification
Live Demo →XGBoost model achieving 97.6% accuracy across 7,200 patients for 3 thyroid conditions. SHAP explainability for clinical interpretability. SMOTE oversampling improved minority class recall from 68% to 93%. Published in Springer.
DSA Solutions Repository
Active DSA preparation in Java organized by data structures and algorithms. Covers arrays, trees, graphs, dynamic programming, and more with optimized solutions.
Blockchain Crowdfunding Platform
Decentralized crowdfunding dApp on Ethereum with Solidity smart contracts implementing withdrawal pattern, milestone-based fund release, and automated refunds. React frontend with Web3.js wallet integration. Zero platform fees, fully on-chain governance.
LabConnect — Research Lab Platform
Full-stack platform connecting students with research labs at Northeastern. Role-based authentication, lab discovery with filtering, and application tracking. Built with React, Node.js, Express, and MongoDB.
SkillSync — Peer Learning Platform
Skill-exchange platform where students teach what they know and learn what they need. Smart matching algorithm pairs complementary skills. Session scheduling with availability management.
Research
Published Work
Classification and Diagnosis of Thyroid Disease Using XGBoost and SHAP
XGBoost classification model for 3 thyroid conditions with SHAP-based explainability. Identified TSH, T3, and T4 as primary diagnostic indicators (65% combined importance). Applied SMOTE to address 3:1 class imbalance, improving minority class recall from 68% to 93%.
Technical Skills
Technologies I Work With
Languages
Backend & APIs
Cloud & DevOps
Databases
Data Science & ML
System Design Patterns
Testing & Practices
Education
Where I Studied
Northeastern University — Boston, MA
CGPA: 3.83 / 4.0
Relevant Coursework
Maharashtra Institute of Technology–World Peace University — Pune, India
CGPA: 9.53 / 10
Relevant Coursework
Contact
Let's Connect