UDAY KIRAN KARUSODI

Andhra Pradesh, India · +91 8185897160


Technical Skills

  • Languages: Java, Python, SQL, C/C++
  • Backend: Spring Boot, Spring Data, REST APIs, Django
  • Databases: PostgreSQL, Redis, Elasticsearch, Faiss, Weaviate
  • DevOps/Cloud: Docker, AWS, HAProxy
  • Frontend: Flutter, HTML, CSS
  • Tools: Git, OAuth2, WebSockets

Experience

Member Technical Staff - Zlabs (R&D Division)

Zoho Corporation · Chennai, India

Jul 2023 - Present

  • Improved plagiarism detection in Zoho Writer by 75% and cut inference time by 43% (0.3s to 0.17s) through optimising a Java-Tomcat backend and a containerised FastAPI & Python service for vector generation (quantised inference, dockerised deployment).
  • Built a prompt-driven RAG pipeline combining LLM calls with keyword & semantic search over Faiss and Weaviate; integrated with crawling and SERP APIs for multi-source retrieval.
  • Migrated crawler to hybrid sync/async model using Java Servlet 3.0 AsyncContext & custom executors, enabling synchronous responses while maintaining webhook scalability.
  • Contributed to the Automatic Extraction pipeline to retrieve structured data from web content by combining image plus text classification with fine-tuning as part of the pipeline; improved extraction accuracy by 60%.

Education

B.Tech in Computer Science Engineering

Gayatri Vidya Parishad College of Engineering (Autonomous) · CGPA: 9.22/10

Visakhapatnam, Andhra Pradesh · Jul 2019 - Jul 2023


Projects

DisciFi (Disciplined Finance Application)GitHub

May 2025 - Present

  • Developed a full-stack app combining finance tracking with behavioural nudges (roast & boast) to improve financial discipline.
  • Implemented secure auth, budgets, transactions, and AI-powered insights driven by MCP.
  • Tech: Java, Spring Boot & Ecosystem, Elasticsearch, PostgreSQL, Redis, RabbitMQ, Google GenAI Toolbox (MCP).

Advanced Object Detection in Images (AICTE Hackathon)GitHub

Nov 2021 - Jan 2022

  • Built an automated CAPTCHA-breaking system using Mask R-CNN and grid-based algorithm, achieving a 74 % success rate against Google reCAPTCHA v2 challenges as part of a national hackathon (Manthan-2022).
  • Tech: Python, Mask R-CNN, PyTorch, Detectron, FastAPI, Docker.


Achievements

  • Contributed 1,500+ LOC across 3+ open-source repos, including chunking libraries and AI agent, etc.
  • Finalist, AICTE Hackathon - Manthan 2021.
  • Finalist, Smart India Hackathon 2022.
  • Core Member, Google DSC (GVP-AI) - Technical Content Speaker.