I'm Diola Chryselle Dsouza

AI & ML Systems Researcher | MS in Computer Science @ Santa Clara University

Technical Arsenal

GenAI & LLMs

  • Large Language Models (LLMs)
  • Multi-Agent Systems & Orchestration
  • RAG (Retrieval-Augmented Gen)
  • Prompt Engineering (CoT, Few-Shot)
  • Model Evaluation (Ragas, DeepEval)

Frameworks & Tools

  • LangChain & LangGraph
  • Vector DBs (Pinecone, ChromaDB)
  • Hugging Face Transformers
  • PyTorch & TensorFlow
  • OpenAI & Anthropic APIs

Engineering & Cloud

  • Python (Expert), Go, Java, C++
  • Distributed Systems (Kafka)
  • REST API & Microservices
  • AWS (SageMaker, Lambda)
  • SQL & Data Engineering

Work Experience

Research Assistant – AI & ML Systems

AIM Lab, Santa Clara University (Lab Website) | July 2025 – Present

  • Architected a Multi-Agent GenAI system using Qwen2.5-7B-Instruct-1M and Hugging Face Transformers for high-quality QA extraction from long-form videos.
  • Designed production-grade RAG pipelines incorporating BGE-M3 vector embeddings, semantic chunking, and agentic routing to optimize context retrieval precision.
  • Developed prompt engineering workflows with system prompts, evaluation filters, and iteration loops to improve relevance and reduce hallucinations.
  • Implemented Python-based agents for chunking, routing, and automated evaluation, focusing on minimizing inference latency and maximizing output accuracy.
  • Authored Beyond Factual QA: Mentorship-Oriented Question Answering over Long-Form Multilingual Content, detailing a novel framework for introducing a multi-agent framework that automates theextraction of in-depth mentorship insights from video transcripts, which outperforms single-agent baselines.
  • Publication: Beyond Factual QA: Mentorship-Oriented Question Answering over Long-Form Multilingual Content (2026). Read Paper | View Code

Software Engineer

SCU Frugal Innovation Hub, Santa Clara University (Lab Website) | May 2025 – January 2026

  • Developed production-ready application features with backend integration via REST APIs.
  • Implemented search, multilingual text-to-speech, and quiz workflows with modular, scalable design.
  • Collaborated cross-functionally to translate product requirements into deployable features and debug issues across UI and API layers.
  • View Code

Software Development Engineer

ICICI Lombard GIC Ltd. | August 2022 – August 2024

  • Built and maintained Python and SQL-based data pipelines supporting enterprise analytics platforms.
  • Integrated real-time data feeds and optimized transformations, reducing insight latency by 30%.
  • Worked with distributed data systems and relational databases to ensure reliability and correctness in production.

Technology Intern

ICICI Lombard GIC Ltd. | February 2022 – August 2024

Developed internal tools and dashboards using Power BI, SQL, and Dynamics 365, improving data quality and reducing manual errors by 20%.

Research Intern

Manipal Institute of Technology | July 2021 – August 2021

Conducted a structured literature review on EEG-based emotion recognition, analyzing machine learning techniques and experimental methodologies.

Education

Master of Science in Computer Science

Santa Clara University, CA

Sept 2024 – June 2026

GPA: 3.81 / 4.0

Bachelor of Technology in Electronics and Communication

Manipal Institute of Technology, India

Minor: Fundamentals of Computing

July 2018 – Aug 2022

GPA: 3.84 / 4.0

Key Projects

Newslet: Distributed Pub/Sub

Event-driven system built with Go, Kafka, and AWS utilizing gossip protocol.

View on GitHub →

Deep Learning Knee Arthritis Detection

Computer vision pipeline using InceptionV3 with attention mechanisms in TensorFlow.

View on GitHub →