I'm an AI / ML and Backend Engineer with experience in machine learning, computer vision, backend development, and software integration.

I enjoy exploring technical challenges across different layers of software, from data preparation and model evaluation to asynchronous programming and system design.

• Backend Engineering — Python, API development, asynchronous programming, concurrency, performance optimization, and system integration.
• Machine Learning — Computer vision, model training and fine-tuning, quantitative evaluation, and inference efficiency.
• Data Engineering — Dataset preparation, annotation, transformation, and data quality improvement.
• Software Engineering — Developer tooling, testing workflows, and AI-assisted software development.

My long-term interests center on model autonomy, self-adaptation, and the development of more flexible and evolving learning systems.

I'm particularly interested in:

• Model Autonomy — Exploring how models can develop mechanisms for internal state representation, goal-directed behavior, and autonomous decision-making.
• Self-Adaptation & Continual Learning — Investigating how learning systems can adapt to changing conditions and accumulate knowledge over time.
• Dynamic Neural Architectures — Exploring structural plasticity and networks capable of adapting or evolving their structures.
• Computational Efficiency — Understanding the trade-offs between model capabilities, adaptability, computational cost, and memory requirements.
• Personalized & Ubiquitous Computing — Creating context-aware computing experiences that enable more natural interactions between people, computing systems, and the physical world.
• AI Agents & Developer Tools — Exploring agents, runtimes, context management, and tool integration as potential technologies for supporting adaptive and personalized experiences.

These are areas I hope to explore further, building on my existing experience in machine learning and software engineering.

An academic team project exploring model autonomy through simulated internal emotional states and goal-directed response selection.

Motivated by my interest in model autonomy, this project investigates how a conversational system might represent an internal emotional state of its own and use configurable target states to guide its response selection.

Rather than relying solely on the user's emotional input, the proposed architecture considers the system's simulated internal state and predicted future states when evaluating candidate responses.

My main contributions included:

• Designing the overall system architecture and integration approach.
• Preparing datasets and training or fine-tuning model components.
• Integrating the components into an interactive conversational prototype.

Individual model components were evaluated, and the prototype was tested through preliminary conversations. However, the broader goals of simulated internal emotional states and autonomous response selection were not systematically validated at the full-system level.

Languages & Backend

Python · FastAPI · Asynchronous Programming · Zenoh · Gstreamer · Rust (academic projects)

Machine Learning & Data

PyTorch · Ultralytics · OpenVINO · MLX · Polars · Hugging Face Datasets

Development & Tools

Docker · Linux · Git · AI Coding Agents