MATIJA JANCEC, PhD

Science Manager/Principal Economist/Data Scientist with 9 years of experience in Big Tech. Passionate for working with big data, enjoying combining Machine Learning and Econometric techniques to find novel solutions for complex problems. Experience in various econometrics, causal estiatmes, time series, forecasting, predictions and causal ML methods. ad23tl@r.postjobfree.com, 347/543-5490, New York, NY

WORK EXPERIENCE

Amazon, Principal Economist, July 2022-

-Lead scientist in a team that supports Amazon Retail negotiation with its vendors (negotiating terms for over $XXXB in purchase per year)

Build a model for an optimal response to vendor cost increase requests. Model is in production and will result in $56MM savings in 2023

Building a novel and foundational science framework to measure and optimize usage of incentives and complex incentives for Amazon negotiation with its vendors (estimated impact in 2024 is $100MM+)

Instagram, Data Scientist (IC7), Communication Systems, August 2021- June 2022

-Creating a new analytical framework to measure, evaluate and optimize over 2B+ notifications send daily

-Establishing a systematic data driven approach to i) evaluate projects using ROI ii) identify the most efficient areas for team to focus and iii) facilitate long term planning and provide strategic overview of the IG Notification space. This led to increased user interactions with IG notifications (+74 bps), without hurting other KPIs

Amazon, Senior Econ Manager (Principal level), August 2014 – July 2021

-Managing a multidisciplinary team (6 scientist + a PM) working to optimize various aspects of Prime Video business using advanced econometric and ML techniques

-Building a prototype for a model providing a unifying framework for content and marketing decisions in Prime Video space, founding and leading a team that productionalized an end-to-end model, with key stakeholders in Creative, Marketing and Finance orgs.

The model provided an early estimate of the title success. This estimate is the main analytical input for the show greenlighting decision

The model produced individual customer prediction, as well as customer segment prediction to watch each title

The model gave an overview of content selection health that surfaced areas where we would benefit the most of brining additional content

The same model was used in optimizing marketing/merchandising effort in PV resulting with the increase in efficiency of 2x-4x compared with the previous approach

Large portion of a team’s time was devoted to experimentation and finding new ways to unlock values for our customers

-Inventory Optimization – working on a project that modeled interaction between sales, prices, customer behavior, Amazon supply/logistics network and inventory

-Topline forecasting – part of a team that worked on forecasting revenue and sold units for the next five years

EDUCATION

PhD in Economics, University of Maryland at College Park, 2009 -2014

Best 3rd year paper award, May 2012

Best Lecturer Award, Summer 2013, 2014

MA in Economics, Central European University (CEU), Budapest, 2007 - 09

Stanislav Vidovic prize for best dissertation

BA in Macroeconomics, The Faculty of Economics, The University of Zagreb, 2003 – 08

MA in Computer Science, The Faculty of Electrical Engineering and Computing, The University of Zagreb, 2001 - 06

COMPUTER SKILLS

PYTHON, SPARK, HADOOP, HIVE, SQL, STATA, Mathematica, C#, C++, C

Contact this candidate