I am a Data Scientist and recently completed postgraduation at University of Delhi. This is my personal website where I share articles on my personal projects, maths, machine learning, algorithms, data structures and more.

πŸ“ Delhi, India
πŸ“§ sagarzendex@gmail.com
πŸ™ GitHub: spartan289
πŸ”— LinkedIn: dev-sagarr


πŸ’Ό Experience

Data Scientist Intern β€” Oriserve, India

Jan 2025 – July 2025

  • Designed and fine-tuned a conversational AI agent tailored to business use cases.
  • Delivered enterprise-scale AI solutions for multinational clients (Bajaj Finserv, Tata Capital, Galactic Client Services), increasing customer engagement by 35%.
  • Built an automated Conversation Evaluation Framework using user behavior modeling, reducing manual validation of conversation voice bots by 80%.
  • Finetuned LLaMA 3.1-8B-Instruct model using LoRA to develop a task-specific user simulation model for dialog system evaluation.
  • Developed a Conversation Analysis Tool to generate insights via thematic clustering of conversation transcripts.
  • Conducted R&D on MuseTalk, a real-time high-quality lip-syncing model with Indian demographic fine-tuning for improved lip-sync accuracy.

Software Engineer Intern β€” FOSSEE, IIT Bombay

Mar 2022 – Aug 2022

  • Developed a CLI version of OSDAG, an open-source project, enabling batch report generation via Excel-driven templates.
  • Built a cross-platform installation setup for Windows and Linux environments.

πŸ›  Skills

Software & Libraries: Numpy, Pandas, Scikit-learn, spaCy, PyTorch, Keras, Matplotlib, Seaborn, FastAPI, Beautifulsoup, Plotly, vLLM

Data Science Skills: Python, SQL, Machine Learning, Deep Learning, LLMs, GenAI, AI Agents, Retrieval Augmented Generation (RAG), Data Analysis, Data Visualisation, NLP, Statistics, Linear Algebra


πŸŽ“ Education

M.Sc. Computer Science β€” University of Delhi
Sept 2023 – June 2025

  • Coursework: Deep Learning, Network Science, Advanced Algorithms, Data Mining, Machine Learning, Image Processing
  • Publication (in review): AugAbEx: Way Forward for Extractive Case Summarization, Springer Nature Language Resources and Evaluation

B.Sc. Computer Science (Hons.) β€” University of Delhi
Nov 2020 – May 2023

  • Coursework: Theory of Computation, Algorithms, Computer Architecture, Computer Networks, Operating System, DBMS

πŸš€ Projects

MathSnap: Text/Image β†’ LaTeX Converter (Chrome Extension)

  • Developed a Chrome extension that converts informally written math formula and math expression’s image snippet into LaTeX format and (vice-versa).
  • Designed for students, educators, and researchers to speed up the process of digitizing math equations.
  • Published on the Chrome Web Store for public use.

Sentence Autocompletion

  • Built an autocomplete system with LSTM and N-gram models for next-word prediction.
  • Tools: TensorFlow, LSTM, RNN

πŸ† Achievements & Certifications

  • GATE DA 2023 Qualified β€” AIR 465, Score 661/1000
  • Top 10 Finalist in IndoML 2024’s Datathon Competition; invited to present at the IndoML 2024 Conference