2026manuscript
Forest Ranger: A Distributed Edge-AI Acoustic Sensor Network Architecture for Real-Time Detection of Illegal Deforestation Activity
Open project manuscript / April 2026 / Vellore Institute of Technology
computer science / software engineering / applied ML
Kasinath C A
Integrated M.Tech CSE (Data Science) at VIT Vellore. I build backend systems, data products and applied ML tools, with a bias toward software that survives real workflows.
Recent work includes Delhivery, video intelligence, edge AI, API security and personal knowledge tooling. I work comfortably across Python, FastAPI, TypeScript, SQL and Docker.
Vellore, Tamil Nadu / Kerala, India
now
devlogAug 2026
Sentinel AI
Video intelligence: detection, tracking, event rules and evidence workflows.
Aug 2026
AgriFlow
Agricultural price analysis using weather, geography and market structure.
Jun-Jul 2026
Delhivery
Software development internship across rider-facing UI, backend/data flows and testing.
selected work
all projectsForest Ranger
ESP32 acoustic sensing node that runs a quantised CNN locally to detect chainsaw activity and send compact alerts.
Mind Palace
Bookmark and content-preservation system with search, annotations, link-health checks and archived copies.
Intent-Aware Security
Identity-verification middleware combining ECDSA challenge-response with Isolation Forest behavioural anomaly detection.
Sentinel AI
Video-security software that connects detection and tracking with rules, evidence capture, camera health and operator response.
technical focus
01Backend & APIs
FastAPI, REST, PostgreSQL, Redis, Docker and testing.
02ML & vision
Anomaly detection, detection/tracking and edge inference.
03Data
Python, R, pandas, geospatial analysis and visualisation.
04Workflow
Git, CI, Playwright, documentation and small automation.
research output
research pageworking style
Map the system
Inputs, state, constraints and failure cases first.
Automate repetition
Small utilities where manual handling keeps recurring.
Leave a trail
Commands, tests and notes that make work easy to resume.
Change tools when needed
Use the stack that keeps the problem clear.