Skip to content

Curriculum vitae#

me

I was born in Sofia, Bulgaria (20.11.1975)

Consistently managing multiple projects with tight deadlines comes naturally.

Scientific research

I specialize in optimization, real-time optimal control, and dynamic simulation, with more than 20 years of modeling experience spanning robotics, economic systems, online bidding, missile navigation, automated vehicles, nuclear reactor instrumentation safety, satellite systems etc. I have a strong background in machine learning and advanced data analysis. I am very skilled at problem abstraction, documentation and presentation.

Software development

I am an experienced software developer with a strong background in Python and C++. I have a good working knowledge across a wide range of programming languages, including Scala/Spark (for large-scale data processing), Rust, Racket, Emacs lisp and several domain-specific languages such as Matlab. I have a solid foundation in algorithms and data structures, with a focus on building efficient, maintainable, and scalable systems. I’m most effective in Linux-based environments, building and maintaining development infrastructure, version control workflows and CI pipelines.

Professional experience#

2021.03 - 2024.05: Kelkoo

senior data scientist (science team)

  • machine learning and operations research for online bidding
  • traffic / resource optimization
  • modeling of merchant-publisher interaction (handling of sparse data)
  • multi-objective formulations for margin optimization
  • python
    • development of company standards for code quality and documentation
    • design/maintenance of common libraries for
      • accessing HDFS through its REST API
      • interfacing optimization solvers and tools
      • convenient use of Jupyter notebooks with scala/spark
    • put in place a local PyPi server
  • scala/spark
    • production quality code implementing operations-research-based-solutions to traffic optimization
    • working with HDFS, YARN
  • in charge of scientific and project documentation
    • designed, implemented and put in practice a custom system for result logging and exchange
  • integrated OR tools and other specialized solvers into the standard pipeline (a game-changer)
  • core participant in the adoption and testing of GitLab CI based workflows

2017.09 - 2021.03: ProbaYes

data scientist/software engineer

  • machine learning applied to various domains

    • fault detection
    • image classification (deep learning)
    • deep reinforcement learning for trajectory generation
    • camera calibration

  • causal analysis

  • concurrent HMM for behavior detection
  • graph analysis
  • time series analysis
  • evaluation of fuzzy controllers
  • development of strategies for fast evaluation of boolean functions
    • using techniques employed in compiler optimization
    • application: quantum computing
  • python
    • fast prototyping in the context of many short-term projects (a skill of its own)
    • support and delivery of code for Linux/Windows
    • development of long-term projects with close client collaboration and tight deadlines
  • C++
    • worked as a consultant/contractor on the code base of Valeo
    • design/maintenance of CMake based build systems
  • rust
    • developed a rocket simulator in Rust for MBDA in the context of a reinforcement learning project

Research experience#

2012.10 - 2015.12: INRIA Rhône-Alpes

research scientist (BIPOP team)

  • fast and reliable solution of lexicographic optimization problems applied to robot control
    • implementation of solver in C++
    • applied to real-world robot safety
  • nonlinear model predictive control
  • robust control for safe locomotion
  • minimum-time control
  • simulation of mechanical systems
  • numerical analysis
  • C++
    • implemented an active-set solver for lexicographic least-squares problems
  • Matlab

2007.12 - 2012.09: Örebro University

research scientist/lecturer (Mobile Robotics and Olfaction Lab at AASS)

  • model predictive control applied to walking robots
  • robotic grasping
  • motion planning for nonholonomic vehicles
  • combining task and motion planning for robotic manipulators
  • control of linear & nonlinear systems
  • simulation of multibody systems
  • introduction to optimization

2006.12 - 2007.12: INRIA Rhône-Alpes

research scientist (BIPOP team)

  • model predictive control for biped robot walking motion generation

Education#

2003.04 - 2006.03: Tohoku University

  • Space Robotics Laboratory Department of Aerospace Engineering
  • degree: Ph.D. in aerospace engineering
  • topic: dynamics and control of space manipulators during a satellite capturing operation
  • advisor: Prof. Kazuya Yoshida
  • awards: Scholarship from the Japanese Ministry of Education MONBU-SHYO (2003.04 - 2006.03)

2003.04 - 2006.03: Hirosaki University

1994.09 - 2001.09: Technical University - Sofia

  • Faculty of Automatics
  • degrees: M.S. in robotics (1999.09), B.S. in electrical engineering (1997.09)
  • advisor: Prof. Veselin Pavlov (M.S. thesis)
  • I decided to move to Japan after having spent two years as a doctoral student

Publications#

Link to my publications

Languages#

  • Bulgarian (native)
  • English (fluent)
  • Japanese (conversational)
  • French (conversational)
  • Russian (basic)