Hey there, I'm Sanjay

• Building: Untitled.

• Designed: and implemented an AI-assisted ticket creation feature for Rule4's TFOX platform using Python, Django 4, MySQL, and a managed RAG stack on AWS (Bedrock text embeddings and LLM, OpenSearch vector search).
• Built: end-to-end retrieval and prompt pipelines that scope historical tickets by client and queue, perform vector similarity search, and generate draft ticket fields while meeting a 2.5s P95 end-to-end latency target.

• Developed: and optimized Java backend services in a modular service architecture, improving request handling and server-rendered performance and increasing Lighthouse Performance scores by 30% across core user flows.
• Implemented: headless content delivery by exposing structured content and metadata via REST and GraphQL APIs, enabling analytics and personalization integrations and reducing manual integration effort by 80%.
• Built: a reproducible CI/CD pipeline with GitLab CI + Maven, adding JUnit automated tests, SonarQube quality gates, and Dockerized build environments to reduce configuration drift and improve release cadence by 35%.
• Implemented: request-level observability using structured logs, correlation IDs, and latency/error metrics, adding production dashboards and alerts that reduced incident triage time by 30% through root-cause isolation.

• Developed: a real estate aggregation platform with Express backend and React frontend, integrating 5+ third-party property listing APIs into a unified search experience.
• Implemented: Redis caching and tuned PostgreSQL queries, improving average search response times by 40% across property listings.
• Built: ETL pipelines in Node.js using scheduled jobs (node-cron), extracting, normalizing, and enriching 10K+ property listings to ensure consistent search and filtering.

• Automated: preprocessing of 50K+ student records with Python (Pandas, NumPy), cutting data inconsistencies 18% and standardizing datasets for advanced analytics.
• Built: interactive Tableau dashboards for attendance and exam performance, boosting data accessibility and enabling 30% faster decisions for educators and administrators.

Relevant Coursework: Datacenter Scale Computing, Network Systems, Database Systems, Big Data Architecture, Intro to Computing Security

Relevant Coursework: Computer Networks, Operating Systems, Machine Learning, Data Structures and Algorithms

Built a browser-based container lab platform with real-time terminals, automated grading, and an AI lab assistant for guided hints. Used Next.js, FastAPI, WebSockets, and Docker-in-Docker to provision isolated per-session containers.

Built a cloud security training platform that provisions sandbox environments, runs Stratus-style attack simulations, and generates MITRE ATT&CK-mapped findings with AI-driven remediation reports. Used FastAPI, Celery, Redis, Terraform, and Gemini.

Built an offline-first macOS desktop port of Monkeytype using Tauri, with local storage, custom theming, and zero server dependencies. Designed it as a fully self-contained typing practice experience.

Built a context-aware notification engine in React Native with Expo to deliver intelligent, real-time location-based push alerts. Integrated Firebase for authentication and messaging, background geolocation tracking, and a GCP-hosted ML model via REST APIs.

Engineered an end-to-end forecasting pipeline in Microsoft Fabric, achieving 91% production forecast accuracy. Used T-SQL for data extraction and Python with Pandas and scikit-learn for feature engineering and model development.

Engineered a distributed audio separation service powered by Demucs on GCP. Deployed Kubernetes-orchestrated Docker containers and integrated Redis for job queueing and MinIO for scalable S3-compatible storage.