Headline

Software Engineer at Oracle | Building AI-powered Data Pipelines | Java • Javascript • Python • AI/ML

About

Software Engineer at Oracle building enterprise AI Copilot and data analytics platforms for 400+ global customers. Core contributor to Oracle's AI Copilot — architected backend orchestration across a 539+ Java-file service integrating WebSocket chat, agentic routing across 15 agent config groups, and macro-driven execution connecting 31 Flowable BPMN workflows. Optimized analytics pipelines reducing response latency from 100s to 20s (80% improvement). Building production ML systems on the side — fine-tuned DistilBERT achieving 92.65% accuracy on 50K reviews, built a RAG pipeline processing 10,000+ document chunks with 3-layer caching using Qdrant and Prometheus observability, and an end-to-end churn prediction pipeline (XGBoost, LightGBM, PyTorch) with SHAP explainability. Core Stack: Backend: Java, Spring Boot, REST APIs, WebSocket, Flowable BPMN, FastAPI ML/AI: Python, PyTorch, HuggingFace Transformers, Scikit-learn, LangChain, SHAP, Qdrant Frontend: ReactJS, TypeScript, Oracle JET Infra: Docker, Prometheus, MLflow, CI/CD Passionate about scalable systems, automation, and applying ML to real-world problems. Active in competitive programming, hackathons, and open-source contributions.

Location

Bengaluru

AI stack

claude code

claude

codex

cursor

hermes agent

hugging face

Work experience

Member Technical Staff-1 at Oracle, Jul 2024 - Present · 2 yrs 2 mos, Bengaluru, Karnataka, India, Full-stack AI engineer on the Data Configuration team within FDI Pipeline division, building enterprise data warehouse solutions serving 400+ global customers. Key Contributions: Backend & Data Engineering: Architected the core backend orchestration layer for Oracle's AI Copilot platform, a natural-language driven development assistant. Built WebSocket-based messaging infrastructure coordinating agentic routing, domain-specific (DS) confirmation workflows, and multi-layer state persistence enabling resumable journeys across connection drops and long-running operations. Designed and implemented macro-driven workflow execution architecture integrating 31 Flowable BPMN process definitions with 49 macro/rule configuration files. This config-driven approach enabled business logic extension through macro groups instead of hardcoded Java paths, accelerating new feature development. Implemented hierarchical folder management system with complex parent-child relationships, supporting rename, delete, and cascading operations Frontend Development: Enhanced ReactJS and Oracle JET interfaces improving user workflow efficiency Built custom code editor with folder navigation, search, and real-time validation System Performance & Security: Identified and resolved critical performance bottlenecks reducing query execution time by 50% Implemented comprehensive testing strategies (unit, integration, E2E) improving code quality and reducing production bugs Collaboration & Mentorship: Mentored junior developers on debugging methodologies, testing best practices, and enterprise system design Collaborated with cross-functional teams across US, India, and Europe time zones Led root cause analysis sessions resolving critical production incidents Tech Stack: Java, Spring Boot, JavaScript, ReactJS, Oracle JET, SQL, MCP, Data Pipelines, Git Impact: Delivered enterprise-grade features enhancing data processing, reporting capabilities, and UI experience for Fortune 500 clients

Server Technology Intern at Oracle, May 2023 - Jul 2023 · 3 mos, Remote, Enhanced multi-parameter search in a Cloud Account Administration tool using ReactJS/TypeScript, adding 3+ filtering parameters enabling efficient filtering across 1000+ account records. Implemented comprehensive test coverage using JUnit and Jest, improving query performance, user experience, and component reliability across ReactJS/TypeScript codebase.

Education

Bachelor of Technology - BTech, Computer Science Engineering, National Institute of Technology Rourkela

Student, 10+2, Mother's Public School

Public links

github: https://github.com/Gourab-18