Leading end-to-end automation and data infrastructure for a fast-growing immigration law firm. I design and implement scalable backend systems that eliminate manual processes, improve data integrity, and integrate legal workflows with cloud platforms.

• Partner closely with attorneys and operations to gather requirements and redesign the client intake process, transforming manual workflows into a fully automated, API-driven pipeline.

• Architected an intake automation pipeline integrating Google Forms, Custom OCR processing, and Clio API to automate data population and validation.

• Implemented automation tooling including visa expiry reminders, LLM-generated consultation summaries, AWS S3 disaster recovery backups, and uptime monitoring for production reliability.

• Currently expanding AI-driven document classification and structured data extraction to further reduce manual case processing.

• Engineered a large-scale ETL migration system to move 6,000+ legacy client folders (1TB+) from a local server to Clio Drive, including automated retries, nested folder recreation, and manifest validation.

• Developed a contact deduplication engine using fuzzy matching and normalized email/phone validation, resolving 4,500+ corrupted or duplicate records.

Worked on Computer Vision system for rooftop and obstacle detection and solar panel layout optimization. Focused on dataset engineering, model performance improvements, and backend integration.

• Built and standardized a 5-channel (RGB, DSM, Flux) rooftop detection dataset using Google Solar API imagery, integrating YOLO-SAM with CVAT to semi-automate COCO annotations across 1,100+ images and improve labeling consistency.

• Resolved critical data augmentation and patching issues using sliding window logic and dynamic padding, stabilizing the training pipeline and eliminating duplication errors.

• Increased model accuracy from 82% to 95% by optimizing multi-channel normalization, augmentation strategies, and loss weighting, while supporting backend integration with a PHP Laravel platform to automate solar layout estimation and cost projections.

Developed secure, scalable backend systems and web applications with a focus on performance optimization and user experience.

• Built RESTful APIs using Java Spring Boot with role-based access control (RBAC) and multi-factor authentication (MFA).

• Optimized backend performance to achieve sub-200ms response times using caching, async processing, and WebSocket-based real-time updates.

• Designed and integrated an AI chatbot for FAQs and user support, reducing support queries by ~25%.

• Contributed to full SDLC including requirements gathering, development, testing, and deployment.