Overview

As an AI/ML Intern at OnPoint, I developed advanced AI applications focusing on generative AI, voice processing, and multi-agent systems for production deployment.

My Role

• Position: AI/ML Developer Intern
• Responsibilities: Built multi-agent chatbots, voice AI applications, and generative systems
• Impact: Delivered two major AI projects with real-time processing capabilities

Key Projects

• AI Storytelling Platform – Generative AI system creating personalized stories based on user inputs
• VitalVoice - AI Voice Gym Coach – Voice-to-voice Android app providing real-time workout guidance

The Challenge

Technical Constraints

• Real-Time Processing - Voice interactions required sub-second response times
• Multi-Agent Coordination - Managing gym trainer and nutrition coach agents simultaneously

Success Metrics

• Sub-1000ms voice response latency
• Seamless multi-agent conversation flow

Our Solution

Key Features

AI Storytelling Platform

Problem solved: Dynamic, personalized narrative generation Tech: Google Gemini API with custom prompt engineering for coherent story generation

VitalVoice - AI Voice Gym Coach

Problem solved: Real-time fitness coaching through voice Tech: LiveKit voice streaming + Google Gemini + LangGraph multi-agent system

Results & Impact

Key Achievements

• Sub-1000ms Response Time - Consistent low-latency voice interactions
• Multi-Agent Success - Seamless gym trainer and nutrition coach coordination