Real-time CV Inference API
Deployed YOLOv8 models behind a robust FastAPI backend, containerized with Docker. Serves real-time object detection and segmentation as the core for multiple client-facing apps.
FastAPIDockerYOLOComputer Vision
Available for new projects
Ahmad Makki
Specializing in Computer Vision, NLP, LLM applications, and the MLOps infrastructure that ships them to production.
// Lahore, Punjab, Pakistan
// about
System info
I'm an AI/ML Engineer and Data Scientist with an MS in Data Science from PUCIT and an Electrical Engineering degree from UET Lahore. I work across Python, SQL, Machine Learning, Deep Learning, NLP, Computer Vision, and Generative AI.
On the engineering side I lean on Docker, Kubernetes, and DevOps practices to deploy and scale reliable models. Across Scraperrs, Machine Learning 1 Limited, and Descon I've delivered predictive modeling, NLP, computer vision, and end-to-end ML deployment. My focus is building practical, reliable, scalable AI that solves real problems.
10+
Production projects
2+
Years experience
9+
Certifications
// services
What I can build for you
Available for freelance on Upwork & Fiverr. From a quick proof-of-concept to a fully deployed AI system — here's how I can help.
Computer Vision Solutions
Real-time object detection, tracking, and segmentation (YOLOv8, OpenCV) delivered as production-ready APIs.
- Custom model training
- Real-time inference API
- Edge & cloud deployment
LLM & RAG Applications
Custom chatbots, document Q&A, and generative features powered by GPT, Gemini, and open-source models.
- RAG over your data
- Chatbots & AI assistants
- Prompt engineering
MLOps & Deployment
Containerize, orchestrate, and ship your ML models to production with Docker, Kubernetes, and CI/CD.
- Dockerized services
- Kubernetes autoscaling
- FastAPI backends
AI Automation & Agents
Automate repetitive workflows and build agentic pipelines with n8n that save your team hours every week.
- n8n workflows
- Agentic pipelines
- API integrations
Have a project in mind?
Let's turn your idea into a working AI product. Hire me directly on your preferred platform.
// core competencies
Technical arsenal
AI / ML / Computer Vision
Modeling, training, and shipping intelligent systems.
PythonMachine LearningDeep LearningGenAINLPComputer Vision
MLOps & Deployment
Containerizing, orchestrating, and scaling models in production.
DockerKubernetesCI/CDAWSDevOps
Core Stack & Tools
The everyday tools I build and automate with.
SQLFastAPIGitLinuxStreamlitn8n
Languages & Libraries
- ▹Python, Pandas, NumPy, Scikit-learn
- ▹PyTorch, OpenCV, NLTK, SpaCy
- ▹Matplotlib, Seaborn
- ▹LLMs (GPT-3.5/4, Gemini, Cohere, Mistral, LLaMA 3)
Tools & Technologies
- ▹VS Code, Jupyter, Google Colab
- ▹Git, GitHub, Linux
- ▹Docker, Kubernetes, FastAPI
- ▹n8n, Streamlit
// work history
Experience
AI/ML Engineer
Scraperrs Lab · Full-time
Lahore, Punjab, Pakistan · On-site
- Engineered end-to-end computer vision solutions (YOLO, Roboflow) for real-time object detection in client projects.
- Built and deployed high-performance FastAPI inference services, containerized with Docker for scalable production use.
- Led client coordination and reporting, translating model performance into business value.
- Architected an automated n8n workflow for an Agentic AI system, streamlining data pipelines and model ops.
- Developed conversational chatbots integrating OpenAI Whisper, LLMs, and in-context learning for multimodal interactions.
Data Science Bootcamp Trainee
Machine Learning 1 Limited · Internship
On-site
- Developed and implemented predictive ML models using Scikit-learn, Pandas, and NumPy.
- Executed end-to-end data preprocessing for supervised and unsupervised learning — cleaning, feature extraction, normalization.
AI Trainee
Descon · Full-time
On-site
- Completed intensive training in ML, NLP, and Generative AI — transformers, prompt engineering, and fine-tuning.
- Gained hands-on MLOps experience containerizing and orchestrating models with Docker and Kubernetes.
// featured work
Projects
Production ML systems — from real-time computer vision APIs to RAG pipelines and Kubernetes orchestration.
Smart Conversational Agent — Mental Health
A multi-faceted conversational agent combining rule-based (TF-IDF), retrieval-based (CNN, LSTM), and generative (GPT-3.5, Mistral) models to provide empathetic mental-health support.
PythonNLPLLMsDeep Learning
RAG-based PDF Query System
Built and deployed a RAG system for PDF querying via a Streamlit interface. Containerized with Docker, pushed to Docker Hub, and deployed to a server for robust access.
DockerStreamlitRAGDeployment
K8s Model Orchestration
Orchestrated an ML model deployment on Minikube with NodePort/LoadBalancer access. Engineered replica scaling (1–5 pods) and CPU-based autoscaling under saturation tests.
KubernetesMLOpsMinikube
// testimonials
What clients say
A few words from people I've worked with on Upwork, Fiverr, and beyond.
“Ahmad built our real-time computer vision API and had it deployed on Docker faster than expected. Clear communication and rock-solid results.”
DR
Daniel R.
Startup Founder · Upwork
“He turned our messy PDFs into a smart RAG chatbot that actually answers correctly. Delivered on time and explained everything simply.”
SM
Sara M.
Product Manager · Fiverr
“Great MLOps skills — dockerized our models and set up Kubernetes autoscaling. Our inference costs dropped and reliability went up.”
OK
Omar K.
CTO, SaaS Company · Upwork
“The n8n automation Ahmad built saves my team hours every week. Professional, responsive, and genuinely knows his stuff.”
LP
Lena P.
Operations Lead · Fiverr
“Fine-tuned an NLP model for us and integrated an LLM assistant. Highly technical yet easy to work with — will hire again.”
BA
Bilal A.
AI Researcher · Upwork
← scroll for more →
// live in-browser neural nets
AI Playground
Models run locally in your browser — nothing leaves your device, and nothing loads until you press Start.
Object Detection
COCO-SSD detects everyday objects in real time from your webcam.
Camera off
Standby
Pose Estimation
MoveNet tracks your body keypoints and draws a live skeleton.
Camera off
Standby
Person Segmentation
MediaPipe cuts you out of the background and drops in a neon backdrop.
Camera off
Standby
// try my ai
Resume ↔ JD Match Analyzer
Paste a job description and a resume — my Gemini-powered analyzer estimates the fit. (Same model I integrate into client projects.)
Processed securely server-side. No key is exposed in the browser.
// analysis output
Paste a JD and resume, then click “Analyze match”.
// knowledge base
Education
MS, Data Science
University of the Punjab (PUCIT)
Sep 2023 — Jul 2025 · CGPA 3.5
Focus: Machine Learning, NLP, Computer Vision, MLOps.
B.Sc, Electrical Engineering
University of Engineering & Technology, Lahore
Graduated
Foundation in mathematics, signal processing, and hardware logic.
// continuous learning
Certifications
UD
Docker Mastery: with Kubernetes + Swarm
Udemy — Bret Fisher · Jan 2025
AI
Generative AI with Large Language Models
DeepLearning.AI (Coursera)
AI
NLP with Classification and Vector Spaces
DeepLearning.AI (Coursera)
AI
Deep Learning Specialization
DeepLearning.AI (Coursera)
AI
Prompt Engineering
DeepLearning.AI (Coursera)
AI
LangChain: Chat with Your Data
DeepLearning.AI (Coursera)
IBM
Python for Data Science, AI & Development
IBM (Coursera)
HR
SQL Skills Certification
HackerRank
HR
Python Skills Certification
HackerRank
// writing
Thoughts & logs
// faq
Frequently asked questions
How long does a typical project take?
It depends on scope. A proof-of-concept can be days; a fully deployed AI system usually takes 1–4 weeks. I'll give you a clear timeline before we start.
How do we communicate and track progress?
We can work over Upwork, Fiverr, email, or WhatsApp. I share regular updates and a working demo as early as possible so you're never in the dark.
Do you offer revisions and post-delivery support?
Yes. Every project includes revisions until you're happy, plus a support window after delivery to fix issues and answer questions.
Can you work with my existing codebase or team?
Absolutely. I can integrate with your current stack, follow your conventions, and collaborate with your developers via Git.
How do you handle data privacy and NDAs?
I'm happy to sign an NDA. Your data and code stay confidential, and I follow secure practices (e.g., API keys are never exposed client-side).
What do your services cost?
Pricing depends on scope and complexity. Share your requirements and I'll send a transparent quote — fixed-price or hourly, whichever suits you.
// get in touch
Let's build something
My inbox is always open — whether it's a role, a project, or just to say hi. I'll get back to you.