-8B5CF6?style=for-the-badge&labelColor=1a1a2e){kind=icon}
Full Stack Engineer with strong Java/Spring Boot backend depth and end-to-end React delivery, building secure, cloud-native microservices and internal platforms in a global banking environment, where I own full-stack systems for compliance validation and org-wide notifications — from REST APIs and data models to React UIs — cutting manual review time by ~78% and streamlining incident workflows. I've also shipped LLM/RAG features in production and engineered a Spring Boot semantic-search platform over 4,800+ questions with Docker-sandboxed execution and a containerized CI/CD pipeline.
- Software Engineer (Sep 2025 – Present) — Java/Spring + React + LLM/RAG systems
- Education — PG-DAC (CDAC, Pune) · B.E. Computer Engineering (Savitribai Phule Pune University)
- Contact — omnaphade0902@gmail.com · linkedin.com/in/omnaphade
Open To: Backend / Full-Stack Engineering roles · AI/ML & LLM integration work · Open-source collaboration
Languages
Frontend
Backend & Databases
Cloud, DevOps & Tooling
| Domain | Proficiency | Details |
|---|---|---|
| LLM Integration | Production | GROQ/Ollama-backed services, prompt engineering, a live multi-agent SRE assistant used by 8 teams |
| Retrieval-Augmented Generation | Production | pgvector semantic search over 4,800+ documents, Ollama embeddings, 5 specialized chat modes |
| Human-in-the-Loop Validation | Production | LLM-assisted compliance review, cutting manual review from ~45 to ~10 minutes per batch |
| Applied ML | Working | scikit-learn Random Forest classifier, ~85% accuracy flagging recurring failures |
Job Portal — Hiring Platform
Built a role-based hiring platform covering the full recruitment lifecycle — candidates, recruiters, and admins — as 9 independent Spring Boot microservices behind an API Gateway, with a React frontend and full Docker Compose local stack.
| Stack | Java 21 · Spring Boot · Spring Cloud · Kafka · PostgreSQL · React · Docker |
| Scale | 9 microservices, event-driven notification pipeline, per-service databases |
| Performance | k6 load tests for login → search → apply and notification-read journeys |
| Security | JWT RBAC, Resilience4j circuit breakers, gateway-level per-IP rate limiting |
| Impact | Prometheus + Zipkin + Grafana observability, JaCoCo-covered CI across all services |
| Repository | github.com/OmNaphade/job-portal |
Architected 9 independent microservices (auth, user, job, company, application, notification, API gateway, config server, service registry) with Eureka service discovery, Kafka-based event-driven messaging, and Resilience4j circuit breakers plus a per-IP rate limiter for fault tolerance and abuse protection. Implemented JWT-based RBAC across three roles with per-service PostgreSQL, a React/Vite/Tailwind frontend including a live admin monitoring dashboard, and Dockerized, CI-tested deployment via GitHub Actions — instrumented with Prometheus metrics, Zipkin distributed tracing, Grafana dashboards, JaCoCo coverage, and k6 load tests.
Interview Hub
Built an end-to-end AI interview preparation platform: layered Spring Boot backend, RAG-grounded chat assistant, and a sandboxed multi-language code playground.
| Stack | Java 21 · Spring Boot · Spring Security · Spring Data JPA · PostgreSQL · pgvector · React/Vite/Tailwind · Docker |
| Scale | 4,800+ interview questions across 12 topics, 5 specialized chat modes |
| Performance | Docker-sandboxed code execution across 8 languages, isolated per run |
| Security | JWT/OAuth via Spring Security, no-network sandboxed containers with capped memory/CPU and execution timeouts |
| Impact | GROQ → Ollama fallback keeps the AI service available under provider outages |
| Repository | github.com/OmNaphade/interview-hub |
Engineered end-to-end: a layered Spring Boot backend (controllers/services/repositories, DTOs, global exception handling) with Spring Security JWT/OAuth auth and a React/Vite/Tailwind frontend, containerized with Docker and a GitHub Actions CI pipeline. Built a RAG pipeline (pgvector, Ollama embeddings) grounding 5 specialized chat modes over 4,800+ interview questions, with a GROQ → Ollama fallback AI service, plus a Docker-sandboxed code playground for 8 languages with per-run isolated containers.
Regulation Impact Intelligence Platform
Built during a 24-hour hackathon (4-person team): an AI-assisted pipeline mapping regulatory obligations to internal controls and applications, scoring risk gaps and generating a prioritized remediation roadmap. Earned judges' recognition for technical execution.
| Stack | Python · FastAPI · React/TypeScript · PostgreSQL · pgvector · Docker · Terraform · GCP Cloud Run |
| Scale | Automated RSS/HTTP ingestion of regulatory changes, role-aware endpoints |
| Security | Role-aware API access, Dockerized deployment on GCP Cloud Run via Terraform IaC |
| Impact | Judges' recognition at a 24-hour, 4-person team hackathon |
| Repository | github.com/OmNaphade/compliance-reg-app |
Architected an AI-assisted regulatory-impact pipeline mapping regulation obligations to internal controls and affected applications, scoring risk gaps and generating a prioritized remediation roadmap. Built a FastAPI backend (SQLAlchemy/Alembic, PostgreSQL + pgvector, role-aware endpoints, automated RSS/HTTP ingestion) and a React/TypeScript dashboard, Dockerized with Terraform IaC for GCP Cloud Run. Used pgvector semantic search to match incoming regulations against existing controls, surfacing coverage gaps and auto-generating an audit-ready remediation summary.
Sep 2025 — Present
Own full-stack systems for compliance validation and org-wide notifications in a global banking environment — from REST APIs and data models to React UIs — plus AI/ML features embedded in production workflows.
- Backend: Engineered Java/Spring Boot microservices with LLM-based, human-in-the-loop validation of ServiceNow records against compliance rules, cutting manual review from ~45 to ~10 minutes per batch.
- API & Data: Designed REST API contracts and relational schemas across PostgreSQL, Oracle, and MySQL; authored JUnit/Mockito unit and integration tests to harden service reliability.
- Full-Stack Platform: Led end-to-end delivery of an internal notification platform (React/Vite + Java/Spring Boot) with RBAC, approval workflows, and lifecycle tracking; onboarded 10+ teams handling 100+ daily notifications.
- Frontend: Delivered React/Vite frontends and full-stack features across multiple internal tools, and integrated SSO (Spring Security + LDAP) for secure access.
- AI Integration: Developed feedback/integration REST endpoints, log-source connectors, and prompt logic for a live multi-agent SRE assistant used by 8 teams to surface fixes from historical incidents and speed triage.
- ML: Trained a Random Forest classifier reaching ~85% accuracy on live data to flag recurring failures across 30+ ServiceNow assignment groups.
Skills:
| Recognition | Details |
|---|---|
| Bank on Tech — Selected Presenter (firm-wide) (Jun 2026) | Selected among firm-wide applicants to present an LLM-powered notification platform for releases and high-impact events to senior management — a framework for transparent, real-time, audience-targeted communication with built-in approval workflows and end-to-end lifecycle tracking |
HackerRank
Google Cloud
DeepLearning.AI
One heatmap for everything — GitHub work plus synced LeetCode/Codeforces solves.
current_focus:
building:
- Job Portal — microservices hiring platform
- Interview Hub — RAG-grounded interview prep platform
exploring:
- LangGraph & multi-agent architectures
- Semantic search at scale (pgvector)
open_to:
- Backend / Full-Stack Engineering roles
- AI/ML & LLM integration collaborationBuilding systems that scale, and systems that think.