Hello, I'm

Vikranth Udandarao

AI Researcher & Data Scientist building reliable agentic AI systems for real-world decision-making.

Applied GenAi @ United Airlines · B.Tech CSAI '26 @ IIIT Delhi · AI Research

Enterprise GenAI

Associate Data Scientist at United Airlines, building production-grade RAG, LLM orchestration, and multi-agent systems for aviation workflows.

Research Impact

Work accepted at AAAI, WWW, India HCI, and ECML PKDD 2026.

Recognition

100% Merit Scholar at IIIT Delhi and Jury-Recommended Winner at the Securities Market Hackathon.

Leadership: Previously served as President of BYLD, IIIT Delhi’s software development club, leading student developers and institute-wide portals used daily by 1,000+ students and faculty.

Research Interests:

Multimodal Reasoning LLM Efficiency Agentic Automation Human-AI Interaction
Vikranth Udandarao

Latest News

July 2026
Thrilled to have returned to United Airlines full-time as an Associate Data Scientist in Gurugram!
May 2026
Oral presentation paper "The Enterprise RAG Paradox: An Industrial Study of Governed and Continuous Documentation Pipelines for LLM-Assisted Code Generation" accepted at ECML PKDD 2026! This work was conducted during my internship at United Airlines.
May 2026
Demo paper "FinMCP: A Multi-Agent LLM Framework for Explainable Real-Time Financial Analysis" accepted at ECML PKDD 2026!
Apr 2026
Selected as one of the 2,000 attendees (out of 25,000+ applicants) to attend Y Combinator Startup School India in Bangalore.
Feb 2026
Joined United Airlines as a Generative AI & Data Science Intern in Gurugram!
Jan 2026
Demo paper "Roamify: An Interactive Browser Extension for Real-Time LLM-Based Travel Itinerary Generation" accepted at The Web Conference (WWW) 2026!
Jan 2026
Paper "Let My Crush See Me: Social-Facilitation Prompting for Zero-Shot LLM Performance Gains" accepted at PromptEng workshop @ The Web Conference (WWW) 2026!
Nov 2025
Received the Best Poster Award at India HCI 2025 for our voice interface project "Suno Samjho Bolo"!
Oct 2025
Paper on "Detecting Citation Hallucinations in Large Language Model Outputs" accepted at the AAAI-26 Student Abstract & Poster Program!
Oct 2025
Paper "Digital-Persona QAR" accepted at the ACM IKDD CODS-COMAD 2025 Young Researchers Symposium!
Sep 2025
Recognized as a Jury-Recommended Winner at the Securities Market Hackathon (Global Fintech Fest 2025) out of 872 national teams.
May 2025
Joined SayaCare as an AI Engineer Intern in Noida!

Publications & Research

ECML PKDD 2026 (Oral)

The Enterprise RAG Paradox: An Industrial Study of Governed and Continuous Documentation Pipelines for LLM-Assisted Code Generation

Udandarao V., Sharma J.K.
Accepted at European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) – Industry Track, 2026 (Oral Presentation)

Conducted an industrial study comparing governed and continuously updated RAG pipelines for LLM-assisted version-migration code generation across 30 prompts and 10 software ecosystems. Proposed a context-aware routing strategy based on downstream code quality. This research was conducted during my AI internship at United Airlines.

@inproceedings{udandarao2026enterprise,
  title={The Enterprise RAG Paradox: An Industrial Study of Governed and Continuous Documentation Pipelines for LLM-Assisted Code Generation},
  author={Udandarao, Vikranth and Sharma, Jeetendra Kumar},
  booktitle={European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD)},
  year={2026}
}
ECML PKDD 2026

FinMCP: A Multi-Agent LLM Framework for Explainable Real-Time Financial Analysis

Parmar A., Udandarao V., Shakya A., Hire T., Anand A., Shah R.R., Wang D.Z.
Accepted at European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) – Demo Track, 2026

Developed an MCP-based multi-agent financial analysis system that combines real-time market data, financial documents, brokerage integration, and web research into evidence-linked analytical responses with sub-10-second latency.

@inproceedings{parmar2026finmcp,
  title={FinMCP: A Multi-Agent LLM Framework for Explainable Real-Time Financial Analysis},
  author={Parmar, Akshat and Udandarao, Vikranth and Shakya, Abhay and Hire, Tanmay and Anand, Avinash and Shah, Rajiv Ratn and Wang, Daniel Zhengkui},
  booktitle={European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) Demos},
  year={2026}
}
AAAI 2026

Detecting Citation Hallucinations in Large Language Model Outputs

Misra N., Udandarao V.
Accepted at AAAI Conference on Artificial Intelligence (AAAI-26) Student Abstract & Poster Program, 2026

Proposed a hybrid bibliographic retrieval and LLM-verification pipeline achieving 80% precision for automatic citation hallucination detection in scholarly text.

@inproceedings{misra2026detecting,
  title={Detecting Citation Hallucinations in Large Language Model Outputs (Student Abstract)},
  author={Misra, Nipun and Udandarao, Vikranth},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={40},
  number={48},
  pages={41325--41327},
  year={2026},
  doi={10.1609/aaai.v40i48.42257},
  url={https://doi.org/10.1609/aaai.v40i48.42257}
}
WWW 2026

Roamify: An Interactive Browser Extension for Real-Time LLM-Based Travel Itinerary Generation

Udandarao V., Vairamuthu M., Tiju N.A., Mistry H., Singh A., Kumar D.
Accepted at The Web Conference (WWW-26) – Demo Track, 2026

Developed a browser-integrated travel assistant using fine-tuned T5 and LLaMA models for attraction summarization and personalized itinerary generation.

@inproceedings{10.1145/3774905.3793121,
author = {Udandarao, Vikranth and Vairamuthu, Muthuraj and Tiju, Noel Abraham and Mistry, Harsh and Singh, Armaan and Kumar, Dhruv},
title = {Roamify: An Interactive Browser Extension for Real-Time LLM-Based Travel Itinerary Generation},
year = {2026},
isbn = {9798400723087},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3774905.3793121},
doi = {10.1145/3774905.3793121},
abstract = {We present Roamify, a browser-embedded demonstration system for real-time, web-conditioned travel itinerary generation. Roamify converts live travel blogs and articles into structured multi-day itineraries by integrating large-scale Web scraping, neural summarisation, explicit preference modelling, and in-browser LLM generation. The system continuously retrieves fresh attraction data, summarises it using a finetuned T5 model, and composes coherent itineraries through LLaMA 3, guided by user-controlled sliders that encode historical, natural, cultural, and amusement preferences. Unlike static trip planning tools or chatbot-only assistants, Roamify enables interactive, controllable, and dynamically updated itinerary creation directly within the Chrome environment. This demo showcases a complete Web intelligence pipeline, from live data acquisition to human-in-the-loop LLM prompting, operating entirely on the client side for transparency and low-latency interaction.},
booktitle = {Companion Proceedings of the ACM Web Conference 2026},
pages = {140–143},
numpages = {4},
keywords = {web intelligence, large language models, personalisation, travel recommendation, browser extension, generative ai},
location = {United Arab Emirates},
series = {WWW Companion '26}
}
CODS-COMAD 2025

Digital-Persona QAR: A Dataset and Pipeline for Financial Question–Answer–Reasoning from Unstructured Documents

Hire T., Udandarao V., Shakya A., Parmar A., Anand A., Shah R.R., Wang D.Z.
Accepted at ACM IKDD CODS-COMAD Young Researchers Symposium, 2025

Introduced a reproducible dataset and LoRA/QLoRA-based long-context pipeline for structured QAR extraction and financial reasoning evaluation from unstructured PDFs.

@inproceedings{hire2025digital,
  title={Digital-Persona QAR: A Dataset and Pipeline for Financial Question--Answer--Reasoning from Unstructured Documents},
  author={Hire, Tanmay and Udandarao, Vikranth and Shakya, Abhay and Parmar, Akshat and Anand, Avinash and Shah, Rajiv Ratn and Wang, Daniel Zhengkui},
  booktitle={ACM IKDD CODS-COMAD Young Researchers Symposium},
  year={2025}
}
WWW 2026 Workshop

Let My Crush See Me: Social-Facilitation Prompting for Zero-Shot LLM Performance Gains

Udandarao V.
Accepted at PromptEng Workshop @ The Web Conference (WWW), 2026

Designed a lightweight social-facilitation prompting technique inspired by psychology that improves zero-shot LLM reasoning performance without explicit chain-of-thought instructions.

@inproceedings{udandarao2026crush,
  title={Let My Crush See Me: Social-Facilitation Prompting for Zero-Shot LLM Performance Gains},
  author={Udandarao, Vikranth},
  booktitle={PromptEng Workshop at The Web Conference (WWW)},
  year={2026}
}
India HCI 2025

Suno Samjho Bolo: Rethinking Voice Interfaces for Multilingual Code-Switching and Digital Inclusion in India

Udandarao V., Misra N.
Accepted at India HCI – Posters, Demos & Artworks Track, 2025 (Best Poster Award)

Designed an offline, emotion-aware voice interface supporting regional code-switching and inclusive multilingual interactions for digital inclusion.

@inproceedings{udandarao2025suno,
  title={Suno Samjho Bolo: Rethinking Voice Interfaces for Multilingual Code-Switching and Digital Inclusion in India},
  author={Udandarao, Vikranth and Misra, Nipun},
  booktitle={India HCI Posters, Demos & Artworks Track},
  year={2025}
}

Research & Experience

Associate Data Scientist

United Airlines
Jul 2026 – Present Gurugram, India
  • Returning full-time to design, build, and deploy production-grade Generative AI systems, RAG solutions, and multi-agent orchestration pipelines for enterprise aviation workflows.

Generative AI & Data Science Intern

United Airlines
Feb 2026 – Jun 2026 Gurugram, India
  • Developing production-grade Generative AI systems for enterprise aviation workflows, focusing on scalability, reliability, and low-latency inference.
  • Building agent-based orchestration pipelines and backend services using LangChain, FastAPI, Redis, and AWS, enabling LLM-driven automation.

AI Engineer Intern

SayaCare
May 2025 – Jul 2025 Noida, India
  • Developed an OCR–LLM integration pipeline using EasyOCR + DeepSeek-R1 (7B) within a Node.js/Flask API, quantized for 6GB GPUs, achieving up to 70% latency reduction and 50% manual annotation savings, while analyzing domain trade-offs.
  • Proposed and evaluated a cosine similarity–based embedding retrieval model for pharmaceutical document matching, improving mapping precision by 25% on validation data.
  • Deployed optimized inference under resource constraints (PyTorch, Hugging Face, OpenCV, Ubuntu GPU) for reproducible performance.

Undergraduate Researcher

Multimodal Digital Media Analysis Lab, IIIT Delhi
Dec 2024 – Apr 2025 New Delhi, India

Advisor: Dr. Rajiv Ratn Shah

  • Investigated multi-agent reasoning architectures via the Model Context Protocol (MCP), designing six autonomous agents (Broker, Indian Markets, Global Stocks, Market Research, Deep Web Research, Digital Twin) for financial analysis automation.
  • Developed the Deep Web Research agent (improving factuality by +33% ROUGE-L) and the Digital Twin agent (enhancing reasoning accuracy by +18% while reducing inference latency from 90s to 10s using vLLM).
  • Conducted reproducible evaluations across FinMCP and FinQAR frameworks to benchmark agent reasoning and retrieval–generation trade-offs.

Selected Work

SIT RAG AI Assistant

SIT RAG AI Assistant

Python · Node.js · FastAPI · AWS EC2 · LanceDB

  • Built a multilingual, voice-enabled RAG assistant serving 500+ daily queries for the Singapore Institute of Technology (SIT).
  • Integrated Elevenlabs STT, TTS and Conversational AI, and hybrid retrieval (LanceDB + BM25) with modular microservices across FastAPI and Node.js.
  • Deployed on AWS EC2 with Nginx, PM2, and systemd, achieving 99.9% uptime and resilient auto-restarts.
FireSafe Security System

FireSafe

Python · YOLO11n · Flutter · Flask · Firebase

  • Engineered a real-time CCTV-based safety system detecting fire and humans using YOLO11n (Ultralytics) fine-tuned on D-Fire + Human datasets.
  • Integrated Flutter mobile app, Flask backend, and Firebase Cloud Messaging for instant safety alerts and remote monitoring.
  • Achieved mAP@0.5 = 0.768, mAP@[.5:.95] = 0.446, and 54 ms/frame inference latency on GPU (640×640 input).

Honors & Leadership

President, BYLD

Software Development Club, IIITD

Led 200+ members, managed institute-wide software portals, and organized major hackathons and developer workshops.

Jury-Recommended Winner

Securities Market Hackathon (Global Fintech Fest 2025)

Honored by SEBI, NSDL, CDSL, NSE, and BSE for designing an innovative fintech solution in a competition containing 872 national teams.

100% Merit Scholarship

IIIT Delhi Academic Excellence

Awarded a full tuition waiver throughout the B.Tech program in recognition of outstanding academic performance.

Batch Representative

Student Council, IIITD

Represented 150+ peers in student council discussions, bridging faculty-student communication and driving academic initiatives.

Get In Touch

Contact Information

Feel free to reach out if you want to collaborate on research, discuss agentic AI, or just say hello.

vikranth22570@iiitd.ac.in
IIIT Delhi, New Delhi, India