As a Microsoft-certified Azure AI Engineer with a solid foundation in cloud-based AI solutions and machine learning concepts. I specialize in building AI/ML models and pipelines using Python, TensorFlow, and Keras, and have hands-on experience with natural language processing (NLP) and image recognition tasks.
Currently working on a personal multi-marketplace (i.e. eBay, Kijiji) aggregator AI agent using Python utilizing the open-source Mistral-7B LLM (via Ollama) for all critical decisions. See below for more details about this project.
Backed by 10+ years at a leading multinational semiconductor company, I bring experience in IP development, system verification, and root cause analysis. Having led global teams for 5+ years, I excel in cross-cultural communication, resource planning, and driving efficiency improvements across complex projects.
I once was told “the best engineers put themselves out of a job”. Hire me and I’ll do just that :)
Skills
Experience Level
Language
Work Experience
Education
Qualifications
Industry Experience
Developed and debugged PySpark data processing scripts for large, real-world datasets (NYC Taxi Trips) using Spark DataFrame API. Performed data ingestion, transformation, filtering, and aggregation
- LLM integration (Mistral 7B via Ollama) for intelligent automation
- Python development and modular architecture
- RESTful API integration (eBay and others)
- Prompt engineering
- Automated pricing, shipping, and inventory analysis
- AI-driven decision making in item selection
Project first started out as linear code. Now developing a scalable Python-based autonomous agent for profit analysis and aggregation across multiple e-commerce platforms, including eBay. Architected the solution using modular software design and robust API integration. Leveraging LLM-powered natural language processing to dynamically interpret listing data, automate complex workflow decisions, and optimize product sourcing strategies in real time. Using Perplexity AI as an assistant.
Key skills demonstrated:
This project showcases practical expertise in advanced AI workflows, applying state-of-the-art language models to solve real-world business problems in online retail and profit optimization.
–
Skills: Python (Programming Language) · Multi-agent Systems · Large Language Models (LLM) · Natural Language Processing (NLP) · Prompt Engineering · Artificial Intelligence (AI) · Machine Learning · AI Agents
Built and trained a neural network in TensorFlow/Keras to classify handwritten digits from the MNIST dataset. Initially implemented a basic single-layer perceptron model (logistic regression using a dense output layer) to establish a performance baseline. Later extended the model to a custom convolutional neural network (CNN) for improved accuracy, incorporating layers such as Conv2D, MaxPooling2D, and Dense. Managed data preprocessing, model training, performance evaluation, and custom image testing.
GitHub Repository: https://www.twine.net/signin
–
Skills: Python (Programming Language) · TensorFlow · Data Analysis · Analytic Problem Solving · Machine Learning
Developed a text classification model using TensorFlow/Keras to analyze movie reviews from the IMDB dataset and predict sentiment (positive or negative). Started with a basic dense-layer architecture before advancing to an embedding-based model with text vectorization and padded sequences. Handled end-to-end development, including data preprocessing, model training, evaluation, and prediction on custom text inputs.
GitHub Repository: https://www.twine.net/signin
–
Skills: Python (Programming Language) · TensorFlow · Data Analysis · Analytic Problem Solving · Machine Learning
Developed a multi-class text classifier using TensorFlow/Keras to categorize Reuters news articles into 46 distinct topics. Employed a Conv1D-based architecture with an embedding layer to extract semantic and contextual features from tokenized input sequences. Implemented sequence padding, global max pooling, and a softmax output for multi-class prediction.
The primary focus was on model calibration — evaluating and improving the reliability of predicted confidence scores using techniques such as Expected Calibration Error (ECE), temperature scaling, and isotonic regression. Assessed model performance on both accuracy and confidence alignment across validation and test sets.
GitHub Repository: https://www.twine.net/signin
–
Skills: Python (Programming Language) · TensorFlow · Machine Learning · Analytic Problem Solving · Data Analysis · Calibration
Practical assessment using Microsoft Azure AI suite and Python.
- Deployed a language resource using pre-built models
- Created a conversational language understanding solution
- Created a custom Named Entity Recognition (NER)
Credential ID: 171B5C649EFABAEA
–
Skills: Python (Programming Language) · Microsoft Azure Machine Learning · Natural Language Processing (NLP) · Artificial Intelligence (AI) · Named Entity Recognition (NER)
Practical assessment using Microsoft Azure AI suite to create a knowledge mining solution using custom skillsets and an enrichment pipeline.
Credential ID: A7C7062B69BA1376
–
Skills: Python (Programming Language) · Microsoft Azure Machine Learning · Data Mining · Artificial Intelligence (AI)
Hire Frederick Suen today
To get started post up your job and then invite Frederick Suen to your job.