Palmaf

2025-present

IsHas Logo

IsHas

IsHas is developed specifically for fire safety inspectors, this application replaces time-consuming manual notes with a streamlined digital workflow that allows for instant, one-click report generation. By centralizing records and implementing role-based access, the platform significantly boosts operational efficiency and ensures full audit readiness.

  • Flutter
  • Python
  • Django

Automatic EEG labeling

This application streamlines the neurological review process by automatically detecting and labeling potential epileptic activity in long-term EEG recordings. By filtering out non-event data, the tool enables specialists to shift their focus from searching for anomalies to performing high-value clinical analysis. See demo here.

  • Python
  • PyTorch
Tabulara Logo

Tabulara

Tabulara.ai is a natural-language data interface that allows non-technical users to extract meaningful insights from raw tabular exports using simple, human-readable commands. By automating the data-processing layer, it transforms "static" exports into an accessible resource for immediate, data-informed decision-making.

  • Python
  • Django
  • HTMX
  • LLM

Personal

Personal server

I maintain a self-hosted Linux server environment using Docker to deploy and isolate various web services behind an Nginx reverse proxy. The project focuses on security best practices, including automated SSL management, firewall hardening, and encrypted remote access.

  • Linux
  • Docker
  • Nginx

Data Scientist certification

I am a DataCamp Certified Associate Data Scientist, having demonstrated technical proficiency in SQL data extraction, Python-based statistical analysis, and predictive modeling. This certification validates my ability to manage the full data lifecycle—from raw data cleaning to the communication of business-relevant insights. [View Certificate]

  • Python
  • Pandas
  • Scikit-learn

Kaggle competitions

I am an active participant in Kaggle data science competitions, where I apply advanced feature engineering and machine learning techniques to solve diverse, real-world predictive challenges. This competitive environment allows me to benchmark my analytical skills against a global community and stay at the forefront of modern data science best practices. [My Profile]

  • Pandas
  • Scikit-learn
  • PyTorch

Deymed

2019-2025

EM field simulation

Physical coil prototype making and measurement is expensive. Computation and visualization of EM fields produces near real-word values that reduces the need of physical coil prototypes to just a few samples.

  • Python
  • Numba

Measurement automation

Medical devices and their components are being thoroughly tested. I integrated several simple elements into a complex system designed to collect and visualize precise measurement data.

  • Python
  • Plotly

TMS Neuronavigation Integration & Prototyping

I managed the technical integration of neuronavigation systems for TMS therapy, focusing on API negotiation with third-party vendors and the development of an in-house C++ prototype. This project established a modular framework for cross-platform data synchronization, ensuring more consistent and accurate therapy delivery for patients.

  • C++
  • gRPC
  • HL7

Internal AI Automation Suite

I developed a suite of internal AI services using Python and LangChain to automate invoice processing, ticket summarization, and RAG-based search. By leveraging Ollama and FastAPI, the system provides secure, high-performance local LLM capabilities that significantly reduce manual workloads and enhance data integrity.

  • Python
  • Ollama
  • FastAPI
  • Langchain

Full-Lifecycle Medical Software Engineering

Developing medical-grade software requires balancing technical performance with strict international safety regulations. I played a key role in the full SDL, ensuring every phase—from requirement analysis to deployment—met the highest quality standards. By integrating IEC 62304 and ISO 14971 into the development process. This approach delivered a safe, audit-ready platform.

Core Responsibilities & Impact

  • Full-Lifecycle Management: From initial architecture and coding to automated build scripts and deployment.
  • Regulatory Compliance: Ensured all development activities adhered to IEC 62304 (Life Cycle) and ISO 14971 (Risk Management).
  • V&V and Quality: Conducted formal validation and created test cases to satisfy rigorous QMS requirements.
  • Cybersecurity: Implemented security measures to protect sensitive medical data and ensure system integrity.
  • Maintenance & Support: Bug fixes, feature enhancements, and localization while providing tier-3 support for global users.
  • Delphi
  • Inno Setup
  • TeamViewer
  • Jira
  • EA

FDA-Cleared Medical Software

I led the development of a new medical software application from initial concept through to successful clinical launch and official FDA clearance. The system is now in active production, serving as a high-integrity platform that meets the stringent safety and efficacy standards.

  • Delphi
  • Jira
  • Inno Setup

Hardware Emulation Framework

I proactively developed a hardware emulation framework that allows developers to test software in a virtual environment without the need for physical devices. This initiative eliminated development bottlenecks, reduced hardware costs, and enabled automated testing of complex edge cases that were previously difficult to replicate.

  • Own Initialive
  • Delphi
  • DLL

University

2013-2019

Rule Extraction from neural networks

Thesis focused on making artificial intelligence more understandable. Neural networks are powerful tools for recognizing patterns and making predictions, but their “black box” nature makes it difficult to see how they arrive at decisions. My thesis explored methods for extracting clear, human-readable rules from these complex models.

Link to thesis on ResearchGate

  • Python
  • Tensorflow