Árpád Goretity

Árpád Goretity

Mentor
Rising Codementor
US$12.00
For every 15 mins
ABOUT ME
Systems software developer, ML engineer, data scientist
Systems software developer, ML engineer, data scientist

My Mission

First and foremost, I am a technical problem solver. As such, I am a strong believer in the multidisciplinary approach: one must learn about a domain as much as necessary and possible, in order to provide a correct, maintainable, and robust solution to the challenge at hand.

I find it exceptionally important to work in a way that I can explain my goals, methods, and results to a technical and non-technical audience at the appropriate level, which makes teaching and mentoring an essential part of my toolbox.

My Experience

I worked in several fields and solved many problems, where mentoring, teaching, and public speaking have always been part of the package:

  • I started my career as a self-taught iOS tweak developer in 2009, during the halcyon days of jailbreaking. With a few coworkers, I ended up completely re-architecting iCsekk, the top utility bill payment application in Hungary, finding and explaining several high-impact security & logic bugs to C-level management. This resulted in a decision of a complete rewrite, which involved fixing bugs, improving maintainability, and accommodating some idiosyncrasies of the backend (which we did not control).
  • Learned Python, JavaScript, shell scripting, and become interested in compilers. I eventually implemented my own scripting language, Sparkling. I learned about optimizations, and gained significant hands-on experience with compiler implementation using LLVM. I introduced Sparkling to the general public at Prezi Tech Talks Budapest, then gave a series of talks at Budapest Swift Meetup on implementing a small, typed language, and a presentation on optimizations at Swift London Meetup.
  • Since my sophomore year, I've been mentoring university students formally (as a teaching assistant and lab instructor) as well as informally, in various formal subjects and broader topics around computer engineering, including: C++, Rust, Python, parallel programming and HPC, linear algebra, calculus, machine learning, and dynamical systems.
  • Wrote my Bachelor's thesis in hardware acceleration of a protein structural alignment method. I implemented the algorithm using HLS in C++ on a Xilinx development board, learning the basics of FPGA-based design in the process. I presented the results as a paper for the ECCTD conference in 2017 in Catania, Italy, as well as a follow-up talk at Budapest Science Meetup.
  • Discovered Rust in 2017 and it quickly became my main language. I gained substantial experience with the ecosystem, published a number of libraries, and contributed to the Rust Users forum.
  • Graduated Summa Cum Laude with an MSc in Data Science from the University of Padua in 2020. Since then, I have been pursing a hybrid data scientist/software architect career at two Hungarian private medical companies, where my tasks range from biostatistical modeling through Typst-based template generation to bioinformatics consulting. I regularly mentored interns and junior team members to improve their coding, algorithm design, and software maintenance skills. During the Covid-19 pandemic, gave an interview about contact tracking methods and applications in the light of mathematical modeling methods.

Language proficiency

  • Certificate of English proficiency, level C2, Cambridge Assessment
  • Accredited French language exam, level C1, Budapest University of Technology.
  • I picked up conversational-level Italian during my 2 years in Padua.
  • My native language is Hungarian.
Hungarian, Italian, English
Budapest (+02:00)
Joined January 2024
EXPERTISE
9 years experience
6 years experience
10 years experience
10 years experience
13 years experience
10 years experience
2 years experience

REVIEWS FROM CLIENTS

Árpád's profile has been carefully vetted and approved as a Codementor. Connect with Árpád now, and leave a review for them once you're done!
SOCIAL PRESENCE
GitHub
Sparkling
Lightweight extension language
C
230
21
libsprec
C library for speech recognition using the Google Speech API
C
134
40
EMPLOYMENTS
Lead Bioinformatics Scientist
Medipredict
2020-01-01-Present

At Medipredict, I am leading a small bioinformatics team of 4 people. Our responsibilities include:

- development and maintenanc...

At Medipredict, I am leading a small bioinformatics team of 4 people. Our responsibilities include:

- development and maintenance of internal tools used by research scientists as well as report generation for our clients;

- making open-source, often research-grade softwer suitable for production use (e.g., via containerization or rigorous re-implementation using high-performance technologies, primarily Rust and Python)

- Mediation between the external IT team and microbiologists, MDs and chemical engineers, to ensure consistency and flexibility architecture of internal databases

- Analysis, visualization, and interpretation of multi-omic data, training and deployment of predictive disease risk models

Python
SQL
Git
View more
Python
SQL
Git
Linux
Bash
Data Science
Rust
Bioinformatics
High-Performance Computing
Biostatistics
View more
Data Platform Architect
S+H Portfolio PLC
2020-01-01-Present
  • Design and implementation of the automated analysis and diagnostic framework of the “PrivátDoktor” personalized telemedicine plat...
  • Design and implementation of the automated analysis and diagnostic framework of the “PrivátDoktor” personalized telemedicine platform from scratch. This software suite helps GPs focus on suspicious cases and obviates the need for continuous human attendance on part of the physician on call.
  • Performance optimization of machine learning models prototyped in Python, by rewriting them in Rust and compiling to WebAssembly.
  • Fitting, tuning, and quality control of machine learning models based on ECG (electrocardiogram) and PPG (photoplethysmogram) signals, for the detection of common heart diseases. This task required extensive application of novel feature engineering methods to time series.
  • Development of an unsupervised on-line learning method for creating personalized physiological profiles for patients based on historical measurements and the detection of anomalies in newly-acquired data. Refactoring of inherited signal processing and analytical code for ensuring correctness and improving maintainability.
  • Documentation of existing and freshly-developed machine learning and diagnosis automation methods for meeting regulatory requirements.
Python
Node.js
SQLite
View more
Python
Node.js
SQLite
TypeScript
Docker
Data Science
Rust
RESTful API
Biostatistics
Scikit-learn
View more
Bioinformatics Scientist
Medipredict Ltd.
2020-01-01-Present
  • Authoring tools for streamlining work of omics scientists, for management of big data, preprocessing, QC, and cleaning of raw dat...
  • Authoring tools for streamlining work of omics scientists, for management of big data, preprocessing, QC, and cleaning of raw data files (genetic and metagenomic)
  • Containerization of research-grade software; implementation of novel algorithms from scratch.
  • Analysis of clients’ microbial and metabolomic data via classical statistical methods. Training and evaluation of predictive machine learning models for rare diseases, with special focus on small in vivo sample sizes.
  • Performing code review for other bioinformaticians. Mentoring interns and trainees with regards to software engineering good practice and data analysis methodology.
Python
SQLite
LaTeX
View more
Python
SQLite
LaTeX
Docker
Data Science
Rust
RESTful API
Biostatistics
Pandas numpy
Scikit-learn
View more
PROJECTS
2022
Parsel is a fully-automated parser generator library. It allows users to specify strongly-typed AST nodes, from which derive macros gener...
Parsel is a fully-automated parser generator library. It allows users to specify strongly-typed AST nodes, from which derive macros generate a parser implementation that directly returns instances of those strongly-typed AST node types.
Rust
View more
Rust
View more
SparklingView Project
2013
Sparkling is a simple, dynamically-typed scripting language, syntactically in the C family, implemented as a portable C library (akin to ...
Sparkling is a simple, dynamically-typed scripting language, syntactically in the C family, implemented as a portable C library (akin to Lua). It aims to be strongly-typed (despite being dynamic) and to correct some design mistakes found in popular scripting languages such as JavaScript.
C
View more
C
View more