Sai Teja Girada

Sai Teja Girada

I am a Computer Science undergrad (B.Tech CSE) at M.V.G.R College of Engineering, Visakhapatnam, interested in building efficient and interpretable machine learning systems. I am currently a Research Intern at NIT Jamshedpur working under Prof. Dinesh Kumar on federated learning with multi-agent reinforcement learning.

Previously, I've worked on historical document OCR with vision-language models, GPT-style transformers with KV cache optimization, audio event detection, and text classification with GloVe embeddings.

Reach me at saitejagirada@gmail.com

Experience

NIT Jamshedpur · May 2026 – Present

Research Intern · Advisor: Prof. Dinesh Kumar

Education

M.V.G.R College of Engineering, Visakhapatnam · 2023 – 2027

B.Tech, Computer Science and Engineering · GPA: 7.73/10.00

Courses: Natural Language Processing, Machine Learning, Data Analytics, Big Data

Research Interests

Deep learning model evaluation and interpretability · Efficient fine-tuning and representation learning · Audio signal processing and event detection · Machine learning systems and deployment pipelines · Federated learning · Multi-agent reinforcement learning

Open Source

Merged pull requests I've contributed to open source projects.

GitHub

github.com/saitejagirada

Blogs

Understanding the Transformer Encoder
A deep-dive into attention, Q/K/V matrices, positional encoding, and how training works — based on the "Attention Is All You Need" paper.
We’ve Been Building AI Applications the Wrong Way
A perspective on the missing infrastructure layer between AI models and applications, and why knowledge should be treated as shared, reusable infrastructure.

Skills

Languages: Python, Java, SQL, C/C++
Cloud & Tools: Azure, AWS, Git, Azure ML
Libraries: PyTorch, pandas, NumPy, scikit-learn, HuggingFace, NLTK
ML Expertise: Pre-training, fine-tuning, post-training, RL, interpretability, ML systems

Last updated July 2026 · saitejagirada@gmail.com