Skip to content
View timkmecl's full-sized avatar

Highlights

  • Pro

Block or report timkmecl

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
timkmecl/README.md

Hi, I'm Tim Kmecl πŸ‘‹

Architecting AI Systems & Statistical Truth
M.Sc. Student in Computer Science & Mathematics @ University of Ljubljana


πŸš€ High-Impact Highlights

  • Early AI Pioneer: Launched one of the world's first ChatGPT IDE integrations within 7 days of OpenAI's public release. Scaled to 400,000+ installs, recognized by Visual Studio Magazine alongside GitHub Copilot, featured in standalone articles by FreeCodeCamp and GeeksForGeeks, and vetted for acquisition by a YC-backed startup.
  • Supreme Court Expert: Authored statistical forensics for a landmark 2026 case at the Slovene Supreme Court. My math identified and measured geographical voting anomalies. Implemented rigorous Python-based forensic pipelines (Hypothesis testing, Linear Regression).
  • Investigative Data Journalist: Using LLMs and Python to audit national records for Mladina, turning raw data into investigative evidence.
  • Cited Researcher: My software is a reference baseline in ACM/IEEE ASE '23 and recommended in Pharma innovation reports by researchers from Genentech & Roche.

πŸ”¬ Peer-Reviewed Research

  • First Author: Natural Language Inference for Slovene (Slovenőčina 2.0, Aug 2024).
    • Comprehensive study on Natural Language Inference (NLI) for the Slovene language, co-authored with Marko Robnik-Ε ikonja.
    • Benchmarked SloBERTa, SloT5, and GPT-3.5 on logical reasoning tasks, used fine-tuning and cross-dataset generalization.
    • Analyzed the "understanding of reality" vs. "pattern matching" in smaller LLMs vs. GPT-class models.

πŸ› οΈ Technical Toolkit

  • AI/LLM: LLM Fine-Tuning, RAG Architectures, API Integration, Prompt Engineering, Automated Data Auditing, Web Scraping.
  • Software Engineering: TypeScript/JavaScript, React, Node.js, Express, Python, Git.
  • Mathematics & Data Science: Python (SciPy, Statsmodels, NumPy, Pandas), Statistics (Linear Regression, Hypothesis Testing etc.), Monte Carlo Simulations, Machine Learning.
  • Strategic Agility: Rapid Prototyping, Product-Market Fit, Technical Due Diligence, Mathematical Rigor.

I specialize in high-stakes technology at the intersection of AI and Rigorous Mathematics. Open to high-impact opportunities in AI Research & Infra, Data Forensics, and Full Stack Web Dev.

Pinned Loading

  1. chatgpt-vscode chatgpt-vscode Public

    VSCode extension that allows you to use ChatGPT or GPT4 inside the IDE

    TypeScript 347 79

  2. studentska-prehrana-scraper studentska-prehrana-scraper Public

    Web Scraper and analysis of changes in Student Food Prices 2022-23

    Jupyter Notebook 3

  3. complex-plane complex-plane Public

    Web app for visualization of complex-valued functions

    JavaScript 1

  4. wayback-bulk-downloader wayback-bulk-downloader Public

    A powerful Python CLI tool and module for bulk downloading web pages from the Internet Archive's Wayback Machine.

    Python 2 1

  5. slo-covid-analysis slo-covid-analysis Public

    Covid data visualization and analysis for Slovenia, spring 2020

    Jupyter Notebook 1

  6. python-image-generation python-image-generation Public

    Experimenting with image generation using Numpy in Python

    Jupyter Notebook 1