AMAlem Memić
Alem Memić (also written Alem Memic) is a Bosnian data scientist and mathematician from Sarajevo. He competed for Bosnia and Herzegovina at the 2006 International Mathematical Olympiad, winning an Honourable Mention, and earned a PhD in mathematics at the University of Sarajevo in 2016 with a thesis on harmonic function theory. He has since worked as a data scientist and machine-learning practitioner on time-series forecasting, computer vision and NLP in Sarajevo, the United States, Poland and Denmark.
Posts and comments mainly on LinkedIn, concentrated on scientific and machine-learning subjects (physics explainers, ML research threads) rather than company promotion; engagement is infrequent and discussion-oriented.
- Currently a Data Scientist at KAKTI (Grindsted, Denmark), engineering and evaluating deep-learning models for multivariate time-series forecasting of power-plant operational data to optimize energy production schedules — a shift from pure mathematics to industrial ML.2024
- His publication output stops around 2016-2017 — the same window as his PhD completion — and his first data-science roles begin immediately after, suggesting he left academia for industry at PhD completion rather than after a postdoc.2017
- The 2006 Olympiad contestant is the same person as the mathematician with a University of Sarajevo PhD (2016); his dissertation 'On a decomposition theorem in harmonic function theory' under advisor Muharem Avdispahić (MSC 31, potential theory) links the teenage competitor to a later pure-mathematics career.2016
- Career trajectory across four countries and seven data-science/teaching roles (KAKTI, itechton, Nexer Insight, Yewno, Arslan Consulting, Checklens, BPU Holdings, University of Sarajevo); the geographic spread Sarajevo->California->Poland->Denmark suggests he follows opportunity across markets rather than anchoring to one hub.2025
- Represented Bosnia and Herzegovina at the 47th IMO in Ljubljana, Slovenia (12-13 July 2006), scoring 11 points (7 on P1, 3 on P4, 1 on P5) and earning an Honourable Mention; the BIH team finished 41st overall.Jul 2006
- Google Scholar profile shows only 3 lifetime citations, h-index 1, most from the 2014 Turkish Journal paper; his research footprint is small and largely dormant since ~2016, consistent with an exit from academia.2026
- Published in harmoni analysis/Bergman-space theory — 'A reproducing kernel for a Hilbert space related to harmonic Bergman space on a domain outside compact set,' Turkish Journal of Mathematics 38 (2014) — a small but genuine research record.Oct 1, 2014
- The 2006 BIH team included future academics Harun Šiljak and Irfan Glogić alongside him; his own path from IMO Honourable Mention (2006) to harmonic-analysis PhD (2016) to industrial data science (2021+) shows Bosnia's math-olympiad pipeline feeding academia and the private sector alike.Jul 2006
- Active on LinkedIn as a commenter on scientific topics (e.g. particle physics explainers, the foundations of physics), suggesting a continued personal interest in pure science well beyond his applied day job.2026
- Awarded a Fund 'Bošnjaci' scholarship for academic year 2011/12 as a postgraduate student of natural sciences in the Department of Mathematics — recognition of early academic promise while still in graduate study.2012
- Maintains a MathSciNet author record (id 1051810) with one indexed publication (2014), and a ResearchGate profile carrying harmonic-Bergman-kernel papers — residual academic presence despite the industry pivot.2014
- Listed in Fund 'Bošnjaci' 2009/2010 records as a student of the Faculty of Science in Sarajevo, Department of Mathematics, confirming his undergraduate home institution.2009
Engineer and evaluate deep learning models for multivariate time series forecasting of power plant operational data, aiming to optimize energy production schedules.
- Trained rotated object detection models to automatically identify and extract structural elements from architectural drawings. - Delivered a solution that drastically reduced man
- Designed and deployed time series models to forecast part delivery schedules and warehouse demand, improving inventory management and reducing carrying costs. - Engineered novel
- Constructed and backtested deep learning models for stock price prediction based on complex, multivariate financial time series data. - Implemented NLP models to extract and stru
- Built an end-to-end identity verification system using facial detection and optical character recognition (OCR), improving the accuracy and speed of customer onboarding.
- Created a high-performance barcode reader capable of handling varied lighting conditions and orientations. - Designed and implemented a fisheye image correction algorithm to pre-
- Developed sentiment analysis classifiers using word embeddings to analyze customer feedback at scale. - Engineered multi-class facial emotion object detection models for visual
Competition record
Bosnia and Herzegovina · IMO