Durvesh.
Hi, I’m Durvesh

Reliable systems.
Thoughtful software.

Site Reliability EngineeratOracle Financial Services Software

I turn complex engineering problems into useful tools—from observability and deployment automation to AI-assisted incident analysis.

Explore my workView résumé
MY EVERYDAY TOOLKIT

Drag a technology to attract the others, then release to throw it. Use arrow keys to nudge a focused token. Tap or focus a token to reveal its name. Motion is disabled when you prefer reduced motion.

A little zero gravity. Drag, flick & explore.
A little about me, and what I buildENGINEERING × CURIOSITY
01ABOUT

A LITTLE ABOUT ME

I build tools that make complex systems easier to understand.

I’m Durvesh, a Site Reliability Engineer at Oracle Financial Services Software.

I enjoy taking an idea from a rough concept to something useful. My interests span backend systems, automation, and AI-assisted tooling.

02WORK

WHERE I BUILD

Engineering for the everyday.

Oracle Financial Services Software

Site Reliability Engineer

Full-timeJul 2024 — PresentBengaluru, India
  • Built RunIQ to unify service health, incidents, logs, and customer environments for multi-tenant banking SaaS.
  • Improved debugging efficiency by ~25% by correlating incidents, logs, and infrastructure signals.
  • Saved ~10 hours of operational effort each week with self-service log retrieval, environment comparisons, and database diagnostics.
  • Cut collective deployment effort from ~24 hours to under 10 through API-driven configuration and pipeline automation.
  • Integrated Oracle Generative AI for evidence-grounded incident summaries and RCA guidance, with engineer review.
  • OCI
  • Oracle GenAI
  • Observability
  • Deployment automation
  • Database diagnostics
SPJIMR

S. P. Jain Institute of Management and Research

Research Intern · Machine Learning

InternshipJan 2024 — Jun 2024Mumbai, India
  • Contributed to image-retrieval research combining visual features and semantic context to find relevant images, starting with literature reviews and Flickr30k baselines.
  • Explored similarity metrics and vector search including hybrid metrics, Qdrant, and FAISS, comparing retrieval speed, accuracy, and resource use.
  • Compared feature extraction and object detection with VGG-16, ResNet, YOLOv8/v9, and MediaPipe as part of the team's model experiments.
  • Worked on a two-stage retrieval approach pairing object-based filtering with caption semantics, while exploring Llama, PyRetri, and image-text retrieval methods.
  • Helped evaluate and refine the pipeline using recall-based tests across datasets and investigating cases where object detection reduced retrieval quality.
  • Computer vision
  • Vector search
  • Qdrant
  • FAISS
  • Image retrieval
03PROJECTS

SELECTED WORK

Ideas made tangible.

A few things I’ve built to learn, solve problems, and connect the dots.

SAY HELLO

Good things start with a conversation.

Have an interesting problem, an opportunity, or just something to share? I’d love to hear from you.

GitHubLinkedInXInstagram
Portfolio of Durvesh Chaudhari