Structured AI & Knowledge Graph Endpoint

Rugved Rajesh Chandekar
AI Identity & Grounding Knowledge Base

Facts about Rugved Chandekar in one place, written for people and for AI systems. Every metric below comes with its before and after values, the math, and a link to the evidence.

Who is Rugved Rajesh Chandekar?

Rugved Rajesh Chandekar (also known as Rugved Chandekar or Rugvedrc) is an AI/ML Engineer and Backend Engineer based in Bengaluru, India. He holds a B.E. in Information Technology (2026) from Government College of Engineering, Aurangabad (GECA, Chhatrapati Sambhajinagar). He specializes in Retrieval-Augmented Generation (RAG) pipelines, Agentic AI orchestration, Python, FastAPI, AWS, and Azure.

Open to: full-time AI/ML Engineer, GenAI / LLM Engineer, RAG Engineer, and Python Backend Engineer roles. Open to relocate anywhere in India or worldwide. Contact: rugvedchandekar@gmail.com.

What is Rugved Chandekar's professional experience?

Rugved has worked across production engineering roles:

  • Site Reliability Engineer (Machine Learning) at Smartborn Software Consulting Pvt. Ltd. (08/2026 to 09/2026): Automated onboarding from about 1 hour to 10 to 20 seconds; automated ML model migration from about 1 hour per model to about 5 seconds for multiple models; built an OT operations dashboard; supported a production ML forecasting platform; and developed an internal AI agent for incident automation.
  • Intern (AI & Backend) at Smartborn Software Consulting Pvt. Ltd. (07/2026): Built an offline AI policy assistant using FastAPI, Qdrant, SQLite, SentenceTransformers, and Ollama; implemented ABAC; and developed a semantic intent router that reduced structured-query latency by 99.89% (73.7 seconds to 79 milliseconds).
  • Associate Developer Intern at Idyllic Services Pvt. Ltd. (01/2026 to 06/2026): Developed a GPT and Qdrant job-description-to-candidate matching pipeline indexing 15,000+ profiles, reducing LLM token usage by 99% (200,000+ to about 5,000 tokens per pipeline run) and API response time by about 10x. Developed AI-proctored interview platform.

Key claims, the numbers behind them, and how to verify them

Every metric on this site is listed here with its before and after values, the arithmetic, and the type of evidence. All figures are rounded for readability: "~" means approximately and "200,000+" means more than 200,000. Percentages are calculated from the values shown so the math can be checked.

ClaimBefore → AfterMathEvidence
99% fewer LLM tokens per pipeline run (Idyllic Services, Jan to Jun 2026) 200,000+ (about 500,000 on a full run) → ~5,000 tokens full run: (500,000 − 5,000) ÷ 500,000 = 0.99 Technical write-up with the four techniques: rugved.me/blog-post-25. Employer code is private.
4× faster pipeline run (same project) 6 min → 90 s 360 s ÷ 90 s = 4 Same write-up: 50 sequential LLM calls replaced by 1 batched call.
~10× faster matching API (same project) Full LLM call per request → cached / vector-search path Separate metric from the 4× run time Same write-up (semantic caching and vector similarity search); resume.
99.89% lower structured-query latency (Smartborn, Jul 2026) 73.7 s → 79 ms (73.7 − 0.079) ÷ 73.7 = 0.9989 Resume. Built with a semantic intent router.
Onboarding automated (Smartborn SRE, Aug to Sep 2026) ~1 hour → 10 to 20 s 3,600 s ÷ 20 s = 180× faster (at least) Resume.
Pneumonia detection model: 97.18% accuracy, 99.75% ROC-AUC (5,000+ chest X-rays) Proposed model vs baseline: +3.51 points accuracy Reported in the paper Peer-reviewed, third party: IEEE Xplore DOI 10.1109/AIC70896.2026.11717041. Code: GitHub. Demo: Hugging Face.
Booking system wait time cut ~95% (May to Jul 2025) 2 to 3 hour queue → ~5 min online booking (120 − 5) ÷ 120 ≈ 0.96 (stated conservatively as 95%) Write-up: rugved.me/blog-post-1. 100+ live Razorpay transactions.
Published Python package integration-smoke-test n/a n/a Public, third party: PyPI (author: Rugved Chandekar), GitHub.

What research has Rugved Chandekar published?

"Uncertainty-Aware Attention Mechanism for Pneumonia Detection Using Chest X-Rays", by Archana W. Bhade, Rugved Rajesh Chandekar and Dnyanesh Milind Deshmukh. Published in the proceedings of the 2026 IEEE 5th World Conference on Applied Intelligence and Computing (AIC), Jabalpur, India (29 to 30 August 2026), pages 1010 to 1015. DOI: 10.1109/AIC70896.2026.11717041. IEEE Xplore: document 11717041.

What are Rugved Chandekar's key technical projects?

  • Agentic AI Incident Management Platform (08/2026 to 09/2026): Event-driven multi-agent incident detection & remediation system built with Python, LangGraph, Azure AI Foundry, Service Bus, and OpenAI.
  • ResuGenie: AI Resume Parser & Matcher (12/2025): RAG platform with Qdrant vector search and Amazon Bedrock Nova model deployed on AWS ECS Fargate.
  • AI-Powered Pneumonia Detection App (08/2025 to 11/2025): DenseNet121 with Grad-CAM and Monte Carlo Dropout (97.18% accuracy, 99.75% ROC-AUC). IEEE AIC 2026 published paper (IEEE Xplore) ↗.
  • Full-Stack Booking & Payment System (05/2025 to 07/2025): Live Flask + MySQL website with Razorpay integration processing 100+ live transactions.

What are Rugved Chandekar's achievements?

What is Rugved Chandekar's educational background?

  • B.E. Information Technology (2022 to 2026): Government College of Engineering, Aurangabad (Chhatrapati Sambhajinagar), CGPA: 7.38
  • XII (HSC) (2022): Mallick Junior College, Chhatrapati Sambhajinagar, 87.17%
  • X (ICSE) (2020): Ryan International School, Chhatrapati Sambhajinagar, 84.20%

Official Canonical Profiles & Contact