Niklas Pechan

Engineer - Machine Learning, Automation and Platforms at Qrious Limited
  • Claim this Profile
Contact Information
us****@****om
(386) 825-5501
Location
Auckland, Auckland, New Zealand, NZ

Topline Score

Topline score feature will be out soon.

Bio

Generated by
Topline AI

You need to have a working account to view this content.
You need to have a working account to view this content.

Experience

    • New Zealand
    • IT Services and IT Consulting
    • 100 - 200 Employee
    • Engineer - Machine Learning, Automation and Platforms
      • Apr 2023 - Present

    • New Zealand
    • Software Development
    • 700 & Above Employee
    • Applied Scientist
      • Feb 2021 - Present

      Working in the Data team delivering an ML solution from R&D to Production, with the aim of enhancing Xero’s bank reconciliation experience. We are partnered with a product team, with our pod consisting of machine learning engineers and data scientists.In this role I have:- Worked heavily on both ML models, applying robust and collaborative data science practice in the design, implementation, review and communication of the results of any model analysis, experiment or error. Experiments I designed resulted in numerous improvements in model performance, including a change to an extractive approach leading to a 60% increase in coverage.- Spearheaded engineering efforts to improve training times for our custom Tensorflow model, resulting a reduction in training times from ~26 hours to ~2 hours. This involved aggressive refactoring of the model and training loop, and the development of an understanding of best-practices for GPU training in NLP. This enabled much faster iteration and experimentation.- Participated in an active research community by giving presentations on the neural architecture used in our extractive and ranking models. - Collaborated with machine learning engineers to deliver models to production, monitoring online performance to ensure it matched offline results.- Provided buddy mentorship to a new team member for the first 6 months in the team. Show less

    • Data Graduate
      • Feb 2020 - Feb 2021

      The year comprised of four rotations through different teams and roles. In order these were: Data Platform Engineering, Applied Scientist, Data Analyst, Data Applications Engineering. The fast paced nature of moving through each role required flexibility to adapt to each team, their ways of working and business requirements.I worked on a number of projects: - sentiment analysis of customer feedback to understand product acceptance - the prediction of account names to reconcile a statement line to, by harnessing pretrained models & applying transfer learning - the production-isation of the aforementioned model - the modernisation of a number of existing scheduled jobs to use tools such as Airflow Show less

    • New Zealand
    • Software Development
    • 1 - 100 Employee
    • Data Science Intern
      • Nov 2019 - Feb 2020

      Bridging the gap between the finish of my degree and starting full time at Xero, I returned to Quantiful for a few months to build on some of the work started the previous summer. This included: - Continuing the experimentation and model development work on the large demand forecasting models. Bridging the gap between the finish of my degree and starting full time at Xero, I returned to Quantiful for a few months to build on some of the work started the previous summer. This included: - Continuing the experimentation and model development work on the large demand forecasting models.

    • New Zealand
    • Software Development
    • 1 - 100 Employee
    • Data Science Intern
      • Nov 2018 - Feb 2019

      This internship was an opportunity to work on machine learning problems in industry. In this role I: - Worked on a large demand forecasting model founded on time series techniques and analysis. - Helped develop a cloud-based experimentation tool that moved the training and inference of models off local laptops to Amazon EC2. - Delivered a presentation to board members about the benefits of the internship programme, and how it helps both companies and the students. This internship was an opportunity to work on machine learning problems in industry. In this role I: - Worked on a large demand forecasting model founded on time series techniques and analysis. - Helped develop a cloud-based experimentation tool that moved the training and inference of models off local laptops to Amazon EC2. - Delivered a presentation to board members about the benefits of the internship programme, and how it helps both companies and the students.

    • New Zealand
    • Software Development
    • 700 & Above Employee
    • Data Engineer Intern
      • Nov 2017 - Feb 2018

      Summer internship at the end of the second year of my degree. First time working in a software team, with cross-locale team members. Countless important lessons during this time, from public speaking and presenting to massive improvements in the quality of my code. Summer internship at the end of the second year of my degree. First time working in a software team, with cross-locale team members. Countless important lessons during this time, from public speaking and presenting to massive improvements in the quality of my code.

Education

  • The University of Auckland
    Bachelor of Engineering - BE, Engineering Science
    2016 - 2019

Community

You need to have a working account to view this content. Click here to join now