Kavindu Sachinthe

University of Kelaniya, Sri Lanka

• Academic Year: 2022/2023 (Year 1) & 2023/2024 (Year 2)
• Overall GPA: Year 1 GPA: 3.56/4.00 | Year 2 GPA: 2.70/4.00* (Not Complete)
• Total Credits Completed: 77

Dehiattakandiya National School, Ampara

• Scholarship Exam, O/L & A/L
• Bilingual Education & Physical Science Stream (with Information & Communication Technology)

Electronics and Computer Science Club

Lead project execution and infrastructure management for the university CTF platform.

AWS Cloud Club - University of Kelaniya

Sasansa Sansada Foundation (Kelaniya University)

• Contributed to a university group project profiling and collaborating with Sasnaka Sansada Foundation's "Corporate Pillar".
• Worked alongside students from diverse districts studying in Kelaniya.
• Researched and analyzed the Foundation's corporate social responsibility initiatives.

Sasnaka Sansada Foundation

• Supported educational initiatives, including the "Ganitha Sawiya" exam paper and remedial seminar series.
• Contributed to STEAM and IT classroom development.

api.kodegas.com

National Award for AI Future Vision (University Category) • Finalist at National Mobile App & Web Dev Competition

• Developed and deployed an end-to-end RESTful API using a TensorFlow CNN model to analyze images of rice plants, providing instant disease diagnosis and treatment.
• Engineered a robust backend with FastAPI and Python, deploying on Microsoft Azure VM using Nginx and Gunicorn.
• Implemented automated CI/CD pipeline using GitHub Actions to update API docs.

t.me/DailyRicePriceBot

• Built a fully automated n8n workflow that fetches daily rice price PDFs published by the Central Bank of Sri Lanka, extracts structured pricing data using Google Gemini AI, and distributes updates.
• Implemented multi-channel distribution including customized HTML emails, a Telegram bot (@DailyRicePriceBot), Google Sheets integration, and MongoDB.

• Built a Retrieval Augmented Generation (RAG) chatbot using Amazon Bedrock Knowledge Base, trained on custom documents including Advanced Level Biology resource books.
• Configured a vector search pipeline using Titan Text Embeddings v2, OpenSearch Serverless, and Amazon S3.
• Deployed Meta's Llama 3.3 70B Instruct model for natural language generation.

cd-kavindu.vercel.app • github.com/kavix/cd-kavindu

• Built a full-stack PWA for real-time home energy monitoring, displaying live power, voltage, temperature, and humidity data from IoT sensors with MongoDB updates every 5 seconds.
• Integrated a time-series predictive analytics model to forecast energy consumption and generate PDF reports.
• Secured via Clerk authentication with middleware route protection.

royalcreditrecoveries.vercel.app • github.com/kavindus0/quizapp

• Developed a comprehensive compliance training platform for Royal Credit Recoveries (PVT) Ltd (PCI DSS, social engineering).
• Role-Based Access Control (RBAC) using Clerk, custom middleware, and 2FA configuration.
• Interactive JWT Token Explorer and backend with Convex.

github.com/RusiruRandika/Modified_Background_Remove_V2

Contributed the web interface layer to an AI-powered image background removal application.

• Designed and built the full Gradio interface, enabling users to upload local images or supply URLs, preview the AI-segmented output, and download transparent PNG results.
• Consolidated integration contributions on Google Colab notebook configurations for instant public sharing and execution.

github.com/openchoreo/backstage-plugins

Contributed feature enhancement to the Backstage plugin ecosystem (PR #643).

• Created the `feat/clickable-log-components` branch to introduce clickable log component functionalities to improve user navigation through build and trace logs.

github.com/thunder-id/thunderid

Contributed to an open-source, high-performance identity stack for managing human and workload access.

• Implemented search and pagination mechanics within the application listing API endpoints to optimize dashboard performance.
• Participated in backend refactoring discussions and issues to decouple `SendOTP()` routines from generic notification modules.

github.com/facebook/pyrefly

Contributed feature enhancement to Facebook's Python project structural analysis framework.

• Submitted a pull request adding support for matching multiple glob patterns within sub-config rules, improving configuration scanning flexibility.