Software, AI and systems built from real problems
Explore software products, AI systems, personal tools, and research projects built by Deependra, including Expense Hive, Pose Estimation, Food Tracker, Outlier, Skift, and Endrare.
Expense Hive is a mobile app built during bachelor's to instantly log expenses, split bills between groups, add reminders, use group chat, and track weekly or monthly spending.
A thesis project using pose-estimation machine learning to analyze athlete movement and provide coaches with real-time body, leg, ankle, and movement attributes for injury-risk awareness.
A personal automation tool for logging food, gym, and daily activity using natural language input, then breaking food entries into nutritional values automatically.
Outlier is an AI system built around context — it learns prompt patterns and interaction history to generate responses that actually fit, cutting the repetitive effort of writing from scratch.
A containerised Flask web application with full end-to-end ownership — from backend development to cloud deployment on AWS.
AWS infrastructure designed with VPC network isolation, EC2 compute, and CloudWatch observability — built and load tested under realistic production conditions.
Skift is a workforce and operations system shaped by real restaurant work, scheduling, communication, sales-aware planning, and AI-assisted decisions.
Endrare is a personal archive for software products, AI projects, case studies, field notes, and product thinking.