Open to opportunities

Sumedha Pandravada

@sumedhapandravada

Message

Backend and AI/ML engineer building LLM RAG and cloud infrastructure for production systems.

India

Message

What I'm looking for

I’m looking for a team where I can build production-grade backend systems and ship real LLM-powered features—RAG, agentic workflows, and reliable cloud infrastructure—while owning end-to-end delivery from system design to deployment and monitoring.

I’m a Software Engineer focused on backend systems, AI/ML, and cloud infrastructure, with hands-on experience across MySQL Server core engineering and MySQL Database Service (OCI) at Oracle.

At Oracle, I fixed severity-2 C++ defects in MySQL Server (including a long-standing production-rooted issue) and backported patches across multiple versions with full MTR validation. I also mitigated 20+ OCI MDS on-call incidents covering OOM, networking faults, and high-availability outages.

I combine engineering rigor with practical AI: I delivered LLM pipelines including Retrieval-Augmented Generation (RAG) and agentic workflows, including an MCP-connected LLM agent in Codex to accelerate MySQL bug triage and root-cause analysis. I also automated 30–40 JIRA tickets/week for faster triage and reduced manual effort.

From internships to projects, I’ve built secure GenAI systems (RAG Q&A with LangChain/FAISS/Azure OpenAI), real-time analytics with Z-score anomaly detection and LLM explanations, multi-agent research pipelines, and production-oriented task queues and full-stack apps deployed on Kubernetes.

Experience

Work history, roles, and key accomplishments

Member of Technical Staff

Oracle

Jul 2025 - Apr 2026 (9 months)

Fixed 3 MySQL Server core C++ sev-2 bugs, including a 9-year-old production defect, by backporting validated patches across MySQL 8.0, 8.4, and trunk. Mitigated 20+ OCI MDS on-call incidents and built a Python automation script processing 30–40 JIRA tickets/week for a 15-member team.

MySQL Server C++Python OCI Distributed Systems CI CD Debugging Model Context Protocol (MCP)

GenAI Intern

Axtria

Jan 2025 - Jul 2025 (6 months)

Built an internal RAG document Q&A system using LangChain, FAISS, and Azure OpenAI, tuned with chunking and top-k retrieval over 50+ documents to improve answer relevance. Developed LLM-powered features including a marketing email generator from Excel/CSV/PDF/TXT inputs and a natural-language data visualization tool that generates and runs Matplotlib code.

RAG Lang Chain Faiss Azure OpenAI Python Natural Language Processing (NLP)Matplotlib Information Retrieval Data Ingestion

Full Stack Intern

Fanztar

Jun 2024 - Dec 2024 (6 months)

Reduced content creation time by ~67% by integrating a GrapeJS visual editor, automating PDF generation, and building a boto3 pipeline to upload assets to Amazon S3. Persisted generated URLs in MongoDB to streamline end-to-end publishing workflows.

Python Amazon S3 MongoDB REST APIs

Education

Degrees, certifications, and relevant coursework

National University of Singapore (NUS)

Academic Research Programme

2023 -

Academic Research Programme (Jul 2023), contributing to a YOLOv8-based safety gear detection model by curating the training dataset and integrating the pipeline into the frontend, achieving 80% precision over 100 epochs.