AMMO AI - AI Content Generation & Marketing Tools

Overview

As a Software Engineer Intern at AMMO AI (Remote, London), I led full-stack development for experimental AI content features and marketing automation tools, shipping production features that increased user engagement and enabled data-driven iteration on AI models.

The Journey

1. Fakers Labs - AI-Generated Podcasts

Problem:

• Experimental AI content lab needed innovative features
• User engagement plateau on existing content
• Needed to showcase AI capabilities creatively
• Complex multi-step content generation required

Solution: Multi-Step LLM Pipeline

Orchestrated Pipeline Architecture:

Existing Content Input
    ↓
Outline Generation (LLM)
    ↓
Script Writing (LLM)
    ↓
Style Transfer (LLM)
    ↓
Text-to-Speech (TTS)
    ↓
AI-Generated Podcast Output

Implementation:

• Led full-stack feature development
• Designed and implemented pipeline orchestration
• Integrated multiple LLM steps sequentially
• Built TTS integration for natural voice output
• Created user interface for podcast generation

Technologies: LLM APIs, Text-to-Speech, Next.js, Python, Pipeline Orchestration

Results:

• 30% increase in user engagement
• Shipped to production on FakersAI
• Demonstrated innovative AI use case
• Smooth multi-step processing

See the result at FakersAI

2. Twitter AI Agent with Analytics Dashboard

Problem:

• Manual social media engagement time-consuming
• Needed to optimize response effectiveness
• No data on what prompts/content work best
• Difficult to measure engagement impact

Solution: Context-Aware AI Agent + A/B Testing Platform

Twitter AI Agent:

• Context-aware response generation
• Integrates with Twitter Dev API
• Sentiment-aware interactions
• Natural, humanized responses
• Supports 1,500 daily interactions

Analytics Dashboard:

• A/B testing framework for prompts
• Engagement measurement (sentiment, click-through)
• Real-time performance metrics
• Data-driven prompt iteration

Technologies: Twitter Dev API, Flask, Next.js, OpenAI API, Analytics Dashboard

Workflow:

Tweet Monitoring
    ↓
Context Analysis
    ↓
Response Generation (A/B variants)
    ↓
Deployment
    ↓
Engagement Tracking
    ↓
Sentiment Analysis
    ↓
Performance Dashboard
    ↓
Prompt Optimization

Results:

• 1,500 daily interactions supported
• Data-driven prompt optimization
• Measured engagement metrics
• Iterative improvement of model performance

3. NeuroMesh Labs Frontend Development

Problem:

• Needed user-facing interfaces for AI Labs features
• Required seamless UI-backend integration
• Had to implement REST API connections

Solution: Developed frontend for multiple Labs pages:

• Text2Image: AI-powered text to image generation interface
• Image2Image: Image transformation and editing tools
• AI Agent: Interactive AI agent interface

Technologies: Next.js, TypeScript, REST API, React

Key Achievements:

• Built 3+ production-ready Labs pages
• Implemented clean UI-backend integration
• Successfully integrated REST API calls
• Collaborated with CTO on UI/UX design

Explore at NeuroMesh

Technical Skills Demonstrated

Full-Stack Development:

• Next.js and React for modern web applications
• TypeScript for type-safe development
• REST API integration
• Flask for Python backend services

LLM & AI Pipeline:

• Multi-step LLM orchestration
• Text-to-Speech integration
• Context-aware AI agents
• Prompt engineering and optimization

Data & Analytics:

• A/B testing frameworks
• Sentiment analysis
• Engagement metrics tracking
• Dashboard development

APIs & Integration:

• Twitter Dev API
• OpenAI API integration
• Webhook implementation
• Real-time data processing

Impact

• Increased user engagement by 30% with AI podcast features
• Enabled 1,500 daily Twitter interactions with AI agent
• Data-driven iteration on prompts and content strategies
• Shipped multiple production features for FakersAI
• Built analytics infrastructure for marketing optimization

Links

• AMMO AI: https://ammoai.io/
• FakersAI: https://fakersai.com/
• NeuroMesh: https://nmesh.io/