Hello, I'm
Aisiri.
MS Computer Science student at Stevens Institute of Technology.
Building agentic AI systems, privacy-first LLM pipelines, and intelligent cloud applications that actually work in production.
New chapter. Same energy.
My Professional Road Map
My Journey
R V College of Engineering
Built a strong foundation in programming, data structures, algorithms, and web development. Graduated with a 7.15/10 GPA. Led cultural organizations and published research.
- –GPA: 7.15 / 10
- –Head of Kannada Sangha — led cultural chapter of 100+ members
- –Cultural Activity Team Secretary (2023–24) — hosted and organized events
- –Published research in IJCRT Vol. 12, Issue 6 (2024)
KodNest
Engineered production-grade HTTP API integrations and end-to-end ML pipelines. Built deep learning models for video classification and anomaly detection.
- –Fault-tolerant API pipelines with retry/backoff logic
- –Deep learning models: PyTorch, TensorFlow, Keras
- –Full ML lifecycle: ingestion → training → deployment
Stevens Institute of Technology
Specializing in machine learning, cloud computing, and advanced algorithms. Building agentic AI systems and privacy-first LLM tools.
- –GPA: 3.89 / 4.0
- –Coursework: Machine Learning, Artificial Intelligence, Cloud Computing (AWS), Advanced Algorithms, Agile Methods for Software Development, Cybersecurity
- –Attended SPY Hackathon — first international hackathon experience
In Motion
Selected Work
A collection of AI/ML engineering projects built for production.
GuardPass
A stateful privacy middleware that intercepts user prompts, detects and anonymizes PII before forwarding to an LLM, then restores original values in the response.
- –Token-mapped PII restoration across multi-turn conversations
- –Pluggable GPT-4o / GPT-4o-mini backends with clean schemas
- –Full-stack: Python backend + React frontend + audit dashboard
AI-Powered Secure Web App
Production-ready, security-conscious web application on AWS combining AI capabilities with cloud-native architecture for scalability and resilience.
- –Deployed on AWS with fault-tolerant, scalable infrastructure
- –Secure authentication and session management
- –End-to-end from frontend to cloud deployment
Video Summarization Storyboard
End-to-end agentic pipeline orchestrating Gemini 1.5, DALL-E 3, and Midjourney as callable tools to generate visual storyboards from YouTube videos.
- –Published in IJCRT Vol. 12, Issue 6, June 2024
- –Orchestrated 3 AI APIs with structured schemas
- –Python ingestion for large-scale unstructured video datasets
Violence Detection in Videos
CNN-based classification model for real-time violence detection in video streams with systematic architecture tuning on large-scale datasets.
- –94% accuracy via systematic CNN architecture tuning
- –Full pipeline: ingestion → preprocessing → feature extraction → eval
- –Reproducible, documented workflows for scalable automation
Railway Trespasser Detection
Real-time ML model for automated trespasser detection using terrain video streams, integrated into a production pipeline for automated action-taking.
- –96% precision on live terrain video streams
- –Production pipeline with automated alert system
- –End-to-end ML in safety-critical visual environments
My Stack
Technical Arsenal
Agentic AI & LLMs
Languages & Frameworks
ML & Deep Learning
Cloud & DevOps
API & Integration
Data & Storage
Get in Touch
Ready to Build the Future?
Active Lifestyle & Hobbies
When I'm not behind the laptop, I'm usually out moving or exploring. Running and hiking/trekking are my favourite reset buttons. I'm a travel freak always hunting for the next new place — and when I'm not on the road, you'll find me on the badminton court, the basketball court, or at the gym. If you share any of these passions, I'd love to connect for a session or a chat!