Hi, I'm Ahmed ML Engineer

3 years shipping generative AI to production — from music generation to mental wellness platforms, currently serving 10K+ users across 5 deployed models.

Cairo, Egypt — Open to Remote & Relocation
Ahmed Khaled Ali - ML Engineer

About

I'm an ML engineer with 3 years of experience deploying generative AI models to production, serving 10K+ users across multiple platforms.

While completing my engineering degree, I built production systems as a sole developer, shipping 5 production models across music generation, LLM inference, and video synthesis platforms.

My unique journey gives me a full-stack perspective — combining web development foundations with cutting-edge ML expertise. I've delivered measurable impact: platforms serving real users, AI models in production, and inference costs optimized at scale.

Focus Areas

GenAI, Production ML, Inference Optimization, LLM Applications, MLOps

Location

Cairo, Egypt — Open to Remote & Relocation

Education

Bachelor of Engineering

Computer & Control Engineering

Obour High Institute for Engineering and Technology Sep 2019 — Jul 2024
Certified

Data Scientist

DataCamp Professional Certificate

+ 5 More Certifications

Experience

Full-Stack ML Engineer

SongLabAI Inc.
Sep 2024 — Present Remote

Sole developer shipping production AI platforms across music generation and mental wellness.

SongLabAI

AI Music Generation Platform
  • Deployed 3 generative AI models into SongLabAI (50+ daily users): Pyramid Flow for video jingle generation, ACE-Step for full song synthesis with lyrics, and fine-tuned MusicGen-Large 3.3B for instrumental generation via Hugging Face Inference Endpoints
  • Reduced inference costs across 5 production models through request batching, response caching, endpoint auto-scaling, and A/B testing deployment configurations
  • Enabled generation of 100+ AI songs across video jingles, full tracks with lyrics, and instrumentals
Hugging Face Endpoints Python WordPress PHP MySQL

AdviceBuddy

AI Mental Wellness Platform
  • Built AdviceBuddy.ai, an AI mental wellness platform reaching 10K+ users and 50K+ conversations, deploying Llama 3.1-8B on Modal GPUs with therapeutic safety monitoring and content filtering
  • Integrated MuseTalk lip-sync model for real-time AI video avatar responses and built text-to-speech pipeline supporting 12 voice configurations (gender × accent) as a premium feature
  • Implemented 3-tier Stripe subscription with server-side rate limiting and transactional emails
Llama 3.1-8B Modal GPUs MuseTalk Content Filtering Stripe

Founding Engineer

Connectyed
Apr 2024 — Sep 2024 Remote • Contract

Built MVP for professional matchmaking platform in 5 months as sole technical founder.

  • Developed Laravel REST API backend and Vue.js SPA frontend with real-time features
  • Created recommendation engine using AWS-hosted ML service with collaborative filtering for match suggestions
  • Integrated Zoom/Google Meet APIs and deployed on AWS production infrastructure
  • Delivered full MVP in 5 months with minimal supervision
Vue.js Laravel PHP MySQL AWS

Freelance ML Engineer & Developer

Self-Employed (Upwork)
Mar 2022 — Sep 2024 Remote
  • Trained and deployed ML models (PyTorch, scikit-learn) and built interactive data analysis dashboards (Plotly, Streamlit) for international clients
  • Developed full-stack web applications with Python and FastAPI, delivering 15+ projects end-to-end across diverse industries
Python PyTorch scikit-learn Plotly FastAPI

Projects

Production systems and ML projects

IN
ML
OUT
MLOps

ML Training & Serving Pipeline

End-to-end MLOps pipeline training DLRM on Amazon Reviews dataset with spot instance scheduling (60-90% cost reduction). Batch prediction, top-k recommendation endpoints, cold-start fallback, and 4 REST API endpoints with response caching.

<1ms Latency
4 Pipeline Stages
60-90% Cost Savings
PyTorch Azure ML FastAPI Parquet
The engine runs smoothly
Toyota delivers excellent quality
Sentiment NER Topics
LLM / NLP

NLP Analysis Pipeline

Multi-feature NLP pipeline with sentiment classification (80%+ F1), BERTopic topic clustering, NER with entity grouping, and 4 interactive Plotly dashboards. 7 NLP tasks deployed live on Hugging Face Spaces.

80%+ F1 Score
7 NLP Tasks
4 Dashboards
Hugging Face Transformers BERTopic Plotly
🧠
MLOps / NLP

Enterprise RAG System

Async document processing pipeline with semantic search, recursive text chunking with deduplication and tiktoken token counting, query reranking, RBAC with JWT auth, and rate limiting. Docker Compose multi-service architecture.

1K+ Chunks/min
<100ms Search
4 RBAC Roles
FastAPI ChromaDB Docker PostgreSQL Redis

Tech Stack

Languages & ML

  • Python, SQL, JavaScript
  • PyTorch, TensorFlow, Hugging Face Transformers
  • LangChain, scikit-learn, Pandas, NumPy

Infrastructure & Frameworks

  • Docker, AWS, Azure ML, Modal, Git
  • PostgreSQL, MySQL, ChromaDB, Redis, Supabase
  • FastAPI, Streamlit, Gradio

Practices

  • Model Fine-tuning, Inference Optimization
  • CI/CD, A/B Testing, Vector Databases
  • GPU Scheduling, Safety Monitoring, Prompt Engineering

Web & Full-Stack

  • Next.js, React, Vue.js, HTML/CSS
  • Laravel, PHP, REST APIs, WordPress
  • Stripe, Supabase Auth, Netlify
"
Ahmed has been instrumental in developing three complete platforms for our companies. He built our professional matchmaking platform with Laravel and Vue.js, our AI music generation platform, and most recently our AI mental wellness companion AdviceBuddy. His ability to deliver production-ready systems that exceed expectations is exceptional.

George Page

Principal, SongLabAI Inc.

Burlingame, California

Get in Touch

Currently exploring new opportunities. Have a project in mind or just want to connect? I'd love to hear from you.