Stanley Sujith Nelavala

Profile Photo
GitHub | LinkedIn | Email
About Me
Skills
Certifications
Research & Teaching Experience
Projects
Publications
Resume

View My GitHub Profile

“If you are fortunate, it is better to build a longer table than a taller fence”

Latest CV

I update my CV regularly to reflect my latest research, projects, and academic commitments. If you’d like a deeper understanding of my work and interests, I encourage you to take a look. it offers the most complete picture of what I do and what I care about.

📄 Stanley’s CV (PDF)

About Me

I’m Stanley Sujith Nelavala, a recent graduate of the M.S. in Computer Science (Research Track) program at the University of Minnesota Duluth, where I completed my degree with a 4.0/4.0 GPA. I hold a Bachelor's degree in Computer Science from IIIT D&M Kancheepuram in Chennai, India.

I plan to pursue a Ph.D. in Computer Science in the upcoming Spring/Fall application cycles. Until then, I am seeking opportunities to apply my skills, broaden my research experience, and contribute to impactful projects in both academic and industry settings. Broadly, my goal is to contribute to the design of AI systems that are not only technically robust, but also socially inclusive and cognitively supportive. My research interests lie at the intersection of Social Computing, Human Computer Interaction(HCI) and Artificial Intelligence(AI). I am particularly focused on studying how large language models (LLMs) can be developed to reflect and support diverse societal values, rather than converging toward a singular or monolithic perspective. I am also very interested in exploring how aligning LLMs with principles of human cognition: such as explanation, feedback, and decision support can help reduce cognitive load and enhance human problem-solving.

Over the past two years, I have investigated multimodal misinformation targeting marginalized communities, focusing on how it is perceived, its societal impact, and the factors influencing its spread. As part of this work, I designed and implemented an end-to-end data pipeline - collecting TikTok browsing histories from Trans and Non-Binary participants through secure web applications, integrating surveys via Prolific, and combining this data with large-scale automated scraping. The system managed over 1 million videos and 300,000 authors, enriched with multimodal features through GPU-accelerated frame extraction, OCR, and metadata harvesting. For analysis, I applied transformer-based topic modeling, clustering, and influence scoring to uncover misinformation themes and creator dynamics, and complemented these computational insights with a perception survey linking data to lived experiences. This project taught me how to work across messy modalities with intellectual honesty, connecting technical rigor to social impact.

I welcome connections—reach out to me on LinkedIn or via email to discuss projects, research, or career opportunities.

Skills

Certifications

AWS Solutions Architect – Associate Badge

AWS Certified Solutions Architect – Associate

Amazon Web Services (AWS)
Issued Feb 2025 • Expires Feb 2028

Research and Teaching Experience

  • Graduate Teaching Assistant, University of Minnesota Duluth (Aug 2023 – Dec 2024)
    • Parallel Programming (Fall 2024)
    • Natural Language Processing (Spring 2024 – Fall 2024)
    • Introductory Python Programming & CS I (Fall 2023 – Spring 2024)
  • Project Intern, IIIT D&M Kancheepuram (Oct 2022 - May 2023)
    • Developed a Python parser to transform Verilog HDL into Virtual Reconfigurable Code (VRC)
    • Enabled downstream simulation, synthesis, and formal verification in FPGA‑style environments
  • Data Science Intern, IIT Kanpur (May 2022 - Oct 2022)
    • Analyzed the publicly available HFEA dataset to estimate IVF treatment success rates
    • Identified key clinical factors influencing outcomes and built interpretable decision‑support models
    • Uncovered data patterns and trends to inform patient‑centric recommendations

    • Projects

    Investigating Multi‑Modal Misinformation on Transgender and Non‑Binary Issues - Masters Thesis (Aug 2023 – Aug 2025)

    • Engineered an end-to-end multimodal data pipeline combining secure participant-facing web apps, Prolific integration, and automated TikTok scraping to collect >1M videos and 300k+ unique authors.
    • Designed a relational MySQL schema for large-scale storage of video metadata, author statistics, and multimodal features with automated enrichment pipelines.
    • Implemented GPU-accelerated frame extraction (FFmpeg) and OCR (PaddleOCR) for on-screen text and caption parsing, integrated with metadata harvesting.
    • Developed NLP/ML workflows using transformer-based BERTopic, K-Medoids clustering, and influence scoring to model misinformation themes, creator dynamics, and amplification networks.
    • Linked computational findings with participant perceptions via a survey study, connecting quantitative pipelines to lived experiences.
    Project Website Thesis Draft

    MarketPulseRT – Real-Time Stock Market Data Pipeline (Dec 2024 – Jan 2025)

    • Developed a real-time data pipeline that streams and processes stock market data using Kafka and Python.
    • Deployed Kafka producers/consumers on EC2, streamed to S3, and transformed using AWS Glue.
    • Enabled real-time querying with Athena over the processed data.
    • Designed a scalable architecture for ingesting and analyzing financial datasets.
    GitHub Link

    Fine-Tuning LLaMA 2 on a Telugu Instruction Dataset (Feb 2025 – Mar 2025)

    • Built a fine-tuning pipeline for LLaMA-2-7B-Chat using QLoRA on a Telugu instruction dataset in Google Colab.
    • Used Hugging Face's transformers, peft, and trl libraries with LoRA for efficient fine-tuning.
    • Managed tokenization, training, and merged model deployment to Hugging Face Hub.
    View Colab Notebook GitHub View Project Paper / Report (PDF)

    Bulldog‑On‑Board: University Transition Assistance App (Jan 2024 – May 2024)

    • Created a Flutter-based mobile app to help incoming university students transition smoothly, with a hub for events, resources, and support.
    • Integrated Firebase for real-time authentication, cloud storage, and personalized student dashboards.
    • Designed an intuitive UI with Figma, featuring interactive maps, reminders, and onboarding tips.
    Prototype(Figma) GitHub (Coming Soon)

    Publications

    Resume

    📄 Download Resume (PDF)