Backend architecture
Service boundaries, clear contracts, and transactional workflows designed around real product behavior.
Available for opportunities
I’m Deepak, a software engineer with 5+ years of experience building scalable backend systems, cloud-native applications, and practical AI products with Java, Spring Boot, React, and distributed systems.
Engineering focus
I build secure, scalable applications with Java, Spring Boot, React, and cloud-native tools, focusing on clear architecture, data integrity, and dependable delivery.
Explore GitHubService boundaries, clear contracts, and transactional workflows designed around real product behavior.
Database locking, validation layers, event-driven persistence, and reliable state transitions.
Containerized workloads, Kubernetes routing, managed infrastructure, and deployment-aware design.
Tool-driven agents, streaming generation, schema grounding, and deterministic guardrails.
Selected engineering work
Four case studies showing how architecture choices turn difficult product requirements into reliable workflows.
A distributed AI application builder that streams generated source code, persists project files, and launches isolated React preview environments on Google Kubernetes Engine.
Turning nondeterministic code generation into a persistent, isolated, browser-accessible application preview.
Separated account, workspace, infrastructure, and intelligence responsibilities across services. Kafka carries file events, MinIO stores project files, Redis maps preview routes, and an idle Kubernetes runner pool prepares isolated Vite environments.
A hotel reservation backend designed around concurrency-safe inventory, composable pricing, secure authentication, and Stripe payment workflows.
Preventing two customers from reserving the same room while keeping pricing rules flexible as business conditions change.
Persistent database row locks and transactional boundaries protect inventory. A strategy chain composes base, surge, occupancy, urgency, and holiday pricing without coupling those rules to the booking lifecycle.
A schema-aware AI pipeline that translates natural-language questions into read-only PostgreSQL queries and records whether each attempt was allowed or blocked.
AI can generate plausible SQL that is malformed, destructive, or unrelated to the database schema.
The service grounds prompts with the live schema, parses model output with JSqlParser, permits one SELECT statement only, rejects unsafe output, and records a clear audit trail before read-only execution.
A structured Spring-based workflow that researches, drafts, reviews, enriches, and saves publication-ready Markdown.
A single large prompt makes research quality and editorial decisions difficult to inspect or control.
Specialized research, writing, and review stages create a traceable pipeline, with Tavily research through MCP, a dedicated reviewer, TLDR generation, reading statistics, and YAML front matter.
Professional experience
Enterprise software, secure APIs, production support, cloud delivery, and accessibility across teams in India and the United States.
Download full resumeTested and remediated websites, applications, documents, and instructional media for WCAG 2.1, Section 508, screen-reader compatibility, and accessible structure.
Built Spring Boot modules and secure REST APIs for enterprise finance workflows, integrating PostgreSQL, Oracle, React, Docker, Amazon RDS, CI/CD, and production release support.
Developed Java microservices and layered REST APIs with PostgreSQL, JPA, Hibernate, Spring Security, JWT, JUnit, and Mockito while supporting delivery and root-cause analysis.
Built and maintained enterprise Java/J2EE applications across controller, service, DAO, JSP, and relational data layers.
Education & certification
University of Illinois Springfield
University of Mumbai
CLF-C02