Vita

Curriculum Vitae #

Work Experience #

  • Applied Scientist at AWS AI, Berlin, Germany. (2019 – present)
  • Applied Science Intern, AI Vertical Services at AWS AI Labs, Palo Alto, CA. (Oct ‘18 – Apr. ‘19)
  • Applied Science Intern, Core ML at Amazon, Berlin, Germany. (Nov ‘17 – Feb. ‘18)
  • Contracting data scientist with several consulting shops in Turkey, US and the mid east. Clients include MasterCard, McKinsey&Company and Delta Partners. (2014 – 2017)
  • Software Engineer at Sentio, an award-winning sports analytics startup where we built player tracking technology for football. (2013 – 2014)
  • Senior Analyst at Peppers&Rogers Group. Johannesburg, RSA; Dubai, UAE and Istanbul. I was a data scientist but it wasn’t called that back then. (2010 – 2012)

Education #

  • PhD, Computer Engineering, Bogazici University, Istanbul, Turkey. 2020. Thesis: Fast High-dimensional temporal point processes. Supervisor: Taylan Cemgil.
  • MS, Software Engineering, Bogazici University. 2014.
  • BS, Industrial Engineering (Op. Res.), Bogazici University. 2010.

Some Recent Work #

  • Minorics L, Turkmen AC, Kernert D, Bloebaum P, Callot L, Janzing D. “Testing Granger Non-Causality in Panels with Cross-Sectional Dependencies.” AISTATS. 2022. link
  • Ansari AF, Benidis K, Kurle R, Turkmen AC, Soh H, Smola AJ, Wang B, Januschowski T. “Deep Explicit Duration Switching Models for Time Series.” NeurIPS. 2021. link
  • Shchur O, Turkmen AC, Januschowski T, Gasthaus J, Günnemann S. “Detecting Anomalous Event Sequences with Temporal Point Processes.” NeurIPS. 2021. link
  • Shchur O, Türkmen AC, Januschowski T, Günnemann S. “Neural temporal point processes: A review.” IJCAI. 2021. arXiv
  • Turkmen AC, Januschowski T, Wang Y, Cemgil AT. “Forecasting intermittent and sparse time series: a unified framework via deep renewal processes.” PLOS ONE. 2021. link
  • Turkmen AC, Capan G, Cemgil AT. “Clustering event streams with low rank Hawkes processes.” IEEE Signal Proc. Letters. 2020.
  • Turkmen AC, Wang Y, Januschowski T. “Intermittent demand forecasting with shallow and deep renewal processes,” NeurIPS Workshop on Temporal Point Processes. arXiv
  • Turkmen AC, Wang Y, Smola AJ. “FastPoint: Scalable Deep Point Processes,” ECML/PKDD 2019. (Best DMKD Paper Award) link

Software #

  • Autogluon is a leading AutoML library.
  • GluonTS is a probabilistic time-series library with a focus on neural time series models.
  • hawkeslib features bare-metal implementations of self-exciting point (Hawkes) process likelihoods in C.
  • I’ve been known to pitch into Apache MxNet (incubating).