About Experience Projects Skills Reviews Contact

Machine Learning Engineer

Ahmed
Danjuma
Haruna

Building production-scale AI systems — from RAG pipelines and LLM orchestration to computer vision platforms — for clients across the globe.

Generative AI LLM Systems MLOps Computer Vision RAG Multi-Agent AI Edge AI LLM Inference
View Projects → Get in Touch
Ahmed Danjuma Haruna
Available for hire
4+
Years in AI/ML
92%
RAG Accuracy
60%
Latency Cut
100%
Job Success
scroll
About

The engineer
behind the systems

I'm a Machine Learning Engineer based in Lagos, Nigeria, with 5+ years of experience designing and deploying AI/ML systems for global clients across healthcare, finance, e-commerce, and travel.

My work spans the full ML lifecycle — from architecting Retrieval-Augmented Generation (RAG) systems and multi-agent LLM frameworks, to building production-grade computer vision pipelines and managing cloud infrastructure at scale.

I have a mathematics background from Adamawa State University, complemented by an MITx MicroMasters in Statistics & Data Science and a Stanford Machine Learning Masterclass — giving me strong theoretical foundations alongside hands-on engineering depth.

LinkedIn ↗ GitHub ↗ Fiverr ↗ Upwork ↗

Generative AI & LLMs

LangGraph RAG Systems Prompt Engineering Stable Diffusion GANs ComfyUI Diffusers

ML Frameworks

PyTorch HuggingFace TensorFlow Keras Scikit-learn XGBoost

Cloud & MLOps

AWS Kubernetes Docker SageMaker CI/CD Pinecone pgvector

Edge AI & LLM Inference

llama.cpp TensorRT ONNX Runtime Model Quantization NVIDIA GPUs INT4/INT8 GGUF

Languages & Databases

Python Node.Js R SQL PostgreSQL
Experience

Where I've built
AI at scale

Senior AI Engineer Jun 2024 – Present
NeuronBytes AI LLC · Part-time
  • Improved diagnostic accuracy by 15% across 100K+ medical images using transformer-based architectures. 100K+ images
  • Maintained 99.5% uptime by deploying AI services using Docker and Kubernetes (EKS).
  • Built PostgreSQL-based data models and ETL pipelines improving data retrieval performance.
  • Optimized inference pipelines using batching and vector database integration.
AI Engineer Aug 2025 – Dec 2025
UnlockAI · Dubai, UAE · Contract
  • Deployed a production RAG chatbot serving 500+ daily users achieving 92% query accuracy using Pinecone and AWS inference infrastructure. ↑92% accuracy
  • Reduced response latency by 60% by building a multi-agent orchestration framework using LangGraph with scalable REST APIs.
  • Improved LLM inference speed by 30% through batching, quantization, and optimized transformer serving.
  • Built evaluation pipelines with logging, metrics, and tracing to improve AI system observability.
AI Course Content Developer Jan 2024 – Jun 2024
WorldQuant University
  • Delivered 20+ AI labs to 1,000+ global students covering computer vision from fundamentals to advanced topics. 1K+ students
  • Designed hands-on exercises on CNNs, GANs, Diffusion Models, and real-world vision applications.
  • Created structured learning materials aligned with modern AI engineering practices.
AI Researcher / Developer Jul 2021 – Jan 2023
OREL Vision
  • Improved image generation model performance by 25% by optimizing GAN and transformer architectures.
  • Reduced model training time by 40% using distributed training pipelines. ↓40% train time
  • Built scalable ETL pipelines processing 50K+ image datasets.
Projects

Selected work
& key systems

Edge AI · Computer Vision · 2024
↗

Real-Time Shoplifting Detection System (Edge AI)

End-to-end computer vision system for detecting shoplifting behavior from CCTV streams. Built full pipeline from dataset creation and annotation to real-time edge deployment, processing RTSP camera feeds with batch inference and ultra-low latency alerts.

Real-time 30 FPS
PyTorch MoViNet YOLO TensorRT ONNX OpenCV RTSP Edge AI
  • Designed and curated custom video dataset (10K+ normal vs 640 shoplifting clips), including manual balancing and preprocessing pipelines.
  • Trained action recognition model using MoViNet with optimized batching (50-frame windows) for temporal inference.
  • Reduced inference latency by converting PyTorch → ONNX → TensorRT with FP16/INT8 optimizations.
  • Built RTSP pipeline for live CCTV ingestion with real-time frame buffering and batch inferencing.
  • Deployed on edge device with GPU acceleration for continuous monitoring without cloud dependency.
  • Integrated alerting system via WhatsApp & Telegram APIs for instant incident notifications.
Generative AI · 2025
↗

OneHaircut — AI Hairstyle Transfer System

Cloud-based ComfyUI workflow for AI hairstyle transfer and coloring, using CNN-based facial alignment and embedding extraction deployed on RunPod.

ComfyUI CNN RunPod Image Generation
Multi-Agent LLM · 2025
↗

CHQ Multi-Agent Travel Chatbot

LangGraph multi-agent system integrating external APIs for travel planning. Designed with agent routing and persistent memory framework for contextual conversations.

LangGraph Multi-Agent REST APIs Memory
Medical AI · 2024
↗

Medical Imaging Diagnostic Platform

Transformer-based architecture for analysing 100K+ medical images with 15% accuracy improvement. Containerised with Docker on EKS with 99.5% uptime SLA.

15%
PyTorch Transformers Docker EKS
RAG / Enterprise · 2024
↗

Production RAG Chatbot — UnlockAI

Pinecone-backed RAG system serving 500+ daily users at 92% query accuracy, with 60% latency reduction via LangGraph multi-agent orchestration on AWS.

92%
Pinecone LangGraph AWS RAG
Computer Vision · 2022
↗

GAN-Based Image Generation Pipeline

Optimised GAN and transformer architectures improving generation quality by 25%, with distributed training pipelines cutting model training time by 40%.

GANs PyTorch Distributed Training
Education AI · 2024
↗

AI Labs Curriculum — WorldQuant University

20+ hands-on AI labs reaching 1,000+ global students, covering CNNs, GANs, Diffusion Models, and real-world computer vision applications.

CNNs Diffusion Models Curriculum Design
Skills

Technical depth
& toolchain

Five years of hands-on engineering across the full ML stack — from research and prototyping to production deployment and observability.

Strongest in LLM system design, RAG architectures, and scalable computer vision pipelines.

LLM Systems & RAG95%
Computer Vision92%
PyTorch / HuggingFace90%
MLOps / Cloud (AWS)85%
Generative AI / Diffusion88%
Data Engineering / ETL80%
Credentials

Education &
certifications

🎓

MITx MicroMasters — Statistics & Data Science

Massachusetts Institute of Technology
2022
🤖

Machine Learning Masterclass

Stanford University
2021
📐

BSc Mathematics

Adamawa State University
2018
☁️

AWS Cloud Practitioner

Amazon Web Services
2023
🔬

Deep Learning Specialisation

deeplearning.ai / Coursera
2021
👁️

Computer Vision Nanodegree

Udacity
2022
Social Proof

Client reviews

5✦
5.0 Fiverr
↑W
100% Job Success
★ Featured Recommendation
Prof. Nicholas Cifuentes-Goodbody
Nicholas Cifuentes-Goodbody
Chief Data Scientist
WorldQuant University, USA

Ahmed is a skilled engineer, a diligent worker, and an enthusiastic collaborator. I would recommend him without reservation for any job in the machine learning or AI space.

I worked with Ahmed in 2024 remotely to assist me in building a course on computer vision for WorldQuant University. Ahmed took the lead in writing all of the code for student lessons and projects using PyTorch, Hugging Face, and other essential tools for computer vision. He meshed into our existing team and tech stack without any issues. His work was always high-quality, on time, and on budget. He was always gracious in taking feedback and incorporating changes. Thanks to his work, this course now enrolls thousands of students from around the world every year.

I enthusiastically recommend Ahmed for any job in data science, NLP, ML engineering, and AI engineering. He would be an excellent addition to any team.

🇺🇸 WorldQuant University, USA · 2024 Academic
D
dimitri_bala 🇨🇭 Switzerland
Fiverr

"Perfect seller. Does what he said, is available, goes in depth in the subject to give best results. Communication was great and I plan to do more business with Ahmed since he is really professional. I Highly recommend this seller."

★★★★★
A
artificial666 🇬🇭 Ghana
Fiverr

"Very incredible programmer. He delivered to my expectations. Would get back to order again."

★★★★★
S
sistem713 🇺🇸 United States · Repeat Client
Fiverr

"Excellent work." — returned for multiple AI chatbot and software projects.

★★★★★
E
Enterprise Client (aico-2023-1) Upwork Enterprise · 4.8 ★
Upwork

"Ahmed was helpful, collaborative, and dedicated. Sometimes he got a little lost in the weeds, and I chalk that up to the depth of care in solving the problem."

★★★★☆ 4.8
C
AWS Client Upwork · Web App / AWS Project
Upwork

"Ahmed did a great job on performing various web app development-related coding tasks on Amazon Web Services. We'd be happy to work with this freelancer again."

★★★★★
P
Python Dash Client Upwork · Data Visualization Project
Upwork

"Ahmed was great, extremely helpful, and patient. He is well-versed in Python pandas and understands modeling. He went over and beyond what I needed."

★★★★★
A
Arbab R. | CEO Upwork · Machine Learning · Dec 2022 · Verified
Upwork

"I hired Ahmed for Computer Vision / Data Science / Deep Learning. He is an expert in his skills. Highly Recommended."

★★★★★
F
fostariha 🇩🇪 Germany · Fiverr
Fiverr

"Perfect seller! Thank you for helping me out. Highly recommend."

★★★★★
Contact

Let's build
something together

Open to freelance projects, contract roles, and full-time opportunities. Whether you need a production RAG system, a computer vision pipeline, or a custom LLM application — let's talk.

✉
profadb88@gmail.com
in
LinkedIn Profile
GitHub
5✦
Fiverr Profile
↑W
Upwork Profile

Send a message