PROJECTS

Projects

A focused view of the work Akshay Penumarthi has built across AI, Python, and full stack development.

The selection spans travel planning, gesture-driven drawing, forecasting, mobile usage analysis, and LLM inference research.

Laptop, tablet, and phone on a wooden desk

Each project pairs practical software with a clear technical idea, built to solve a distinct problem.

WORK

Projects built around practical ideas and technical depth.

These projects show how the work shifts between product thinking, machine learning, interface design, and applied development.

AI TRIP PLANNER

AI Trip Planner

An AI-powered travel planning application using Google Gemini, OpenStreetMap Nominatim, Leaflet, and OpenRouteService for destination search, route visualization, travel suggestions, and budget estimation.

Python Flask JavaScript

VIRTUAL DRAWING

Virtual Drawing Using Hand Gestures

A real-time virtual drawing application built with hand tracking and gesture recognition, combining OpenCV, MediaPipe, Tkinter, and Python for a responsive canvas experience.

OpenCV MediaPipe Tkinter

FARMERS RETAIL MART

Farmers Retail Mart

A sales forecasting and inventory management application that analyzes sales data and supports inventory decisions, built with Python, Flask, MySQL, and machine learning.

MySQL Flask Machine Learning

MOBILE ADDICTION

Mobile Addiction Detection

A machine learning project that analyzes mobile usage patterns through data analysis and predictive modeling to surface meaningful behavior signals.

Data Analysis Predictive Modeling

RESEARCH

A project that focuses on how models should be chosen and run.

The research-oriented project centers on inference optimization and the practical tradeoffs between models, quantization methods, and runtimes.

ADAPTIVE LLM INFERENCE ENGINE

Adaptive LLM Inference Engine

A research-oriented project for intelligent LLM inference optimization, with a focus on selecting suitable models, quantization methods, and runtimes based on task requirements and hardware capabilities.

  • ✓Model selection guided by task requirements.
  • ✓Quantization choices matched to hardware limits.
  • ✓Runtime decisions shaped around deployment context.
A focused shot of a hand placing a sticky note on a desktop monitor in an office setting.

CONTACT

Send a message about a project.

If you want to talk through a build, a collaboration, or a technical idea, reach out with a short note and the context you have in mind.