Experience

uxo.ai

• Information Services
• 1 - 100 Employee

• CEO, co-founder

  • Sep 2023 - Present

  Building an AI that autonomously (yet controllably) navigates and extracts information from the web. Building an AI that autonomously (yet controllably) navigates and extracts information from the web.



Sumitomo Mitsui Finance and Leasing Company, Limited

• Financial Services
• 1 - 100 Employee

• Machine Learning Engineer, ML Operations

  • Feb 2021 - Present

  Working inside the Digital division's AI team. Implemented MLOps platform for the data science team (Polyaxon, Kubernetes, AWS, Terraform). Finding ways to increase automation, code quality, and experiment repeatability. A big focus is on removing manual steps from model productionisation, without reducing flexibility in the exploratory work of data scientists. Working with deep learning models for optical character recognition (OCR) using TensorFlow (more use cases to follow). Working inside the Digital division's AI team. Implemented MLOps platform for the data science team (Polyaxon, Kubernetes, AWS, Terraform). Finding ways to increase automation, code quality, and experiment repeatability. A big focus is on removing manual steps from model productionisation, without reducing flexibility in the exploratory work of data scientists. Working with deep learning models for optical character recognition (OCR) using TensorFlow (more use cases to follow).



Subspace AI Ltd

• United Kingdom
• IT Services and IT Consulting

• CEO

  • Feb 2018 - Jun 2021

  Offered services in full stack ML engineering, model development, training, deployment as a contractor. Offered services in full stack ML engineering, model development, training, deployment as a contractor.



ProArch

• United States
• IT Services and IT Consulting
• 200 - 300 Employee

• Machine Learning Engineer

  • Apr 2020 - Jan 2021

  Worked on a distributed multi-stage pipeline using PySpark, Azure Databricks, and various Azure services (including their CI/CD tools). Worked on a distributed multi-stage pipeline using PySpark, Azure Databricks, and various Azure services (including their CI/CD tools).



FunnelAI

• United States
• IT Services and IT Consulting
• 1 - 100 Employee

• Machine Learning Engineer

  • Jun 2019 - Mar 2020

  Built a continuous model retraining, tuning & deployment system on Kubeflow. The design emphasises DRY, reusable components, scalable data processing, and ease of adding additional models and training objectives. Performed multi-objective training on top of SotA NLP models (BERT, T5). Built a continuous model retraining, tuning & deployment system on Kubeflow. The design emphasises DRY, reusable components, scalable data processing, and ease of adding additional models and training objectives. Performed multi-objective training on top of SotA NLP models (BERT, T5).



Factmata

• United Kingdom
• Public Relations and Communications Services
• 1 - 100 Employee

• Machine Learning Engineer

  • Jan 2019 - Jun 2019

  Designed and lead the development of a multi-stage NLP pipeline (on Kubeflow) for a media monitoring tool, involving clustering, semantic similarity, keyphrase extraction, etc. Designed and lead the development of a multi-stage NLP pipeline (on Kubeflow) for a media monitoring tool, involving clustering, semantic similarity, keyphrase extraction, etc.



SHAREIGHT

• United Kingdom
• Advertising Services
• 1 - 100 Employee

• Machine Learning Engineer / Data Scientist

  • Nov 2017 - Nov 2018

  Built an end-to-end ML pipeline for product categorisation and visual similarity. Employed multi-objective, multi-modality, active and transfer learning approaches. Trained the team to use the ML model independently. Used the GCP stack (Dataflow, ML Engine, PubSub, Dataprep, GCS, CloudSQL) and open source software (TensorFlow, Airflow, Apache Beam, ElasticSearch, RabbitMQ, Flask, React). Built an end-to-end ML pipeline for product categorisation and visual similarity. Employed multi-objective, multi-modality, active and transfer learning approaches. Trained the team to use the ML model independently. Used the GCP stack (Dataflow, ML Engine, PubSub, Dataprep, GCS, CloudSQL) and open source software (TensorFlow, Airflow, Apache Beam, ElasticSearch, RabbitMQ, Flask, React).



Apparat Cloud OÜ

• Estonia, United Kingdom

• Founder

  • Jun 2008 - Dec 2017

  Offered services as a machine learning / data science agency. Formerly: contract-based web development for a variety of clients. Offered services as a machine learning / data science agency. Formerly: contract-based web development for a variety of clients.



Lingvist

• Estonia
• E-Learning Providers
• 1 - 100 Employee

• Machine Learning Engineer / Data Scientist

  • Jul 2017 - Nov 2017

  Built a generative / discriminative neural NLP model for semisupervised learning and content generation. Using TensorFlow, tensor2tensor, Keras. Built a generative / discriminative neural NLP model for semisupervised learning and content generation. Using TensorFlow, tensor2tensor, Keras.



Rise Art

• United Kingdom
• Artists and Writers
• 1 - 100 Employee

• Machine Learning Engineer / Data Scientist

  • Jun 2016 - Jun 2017

  Built a recommender engine based on deep learning and matrix factorization using Python, TensorFlow, Spark and Google Cloud ML Engine. Implemented and measured the performance of various recommender algorithms. End-to-end engineering. Built a recommender engine based on deep learning and matrix factorization using Python, TensorFlow, Spark and Google Cloud ML Engine. Implemented and measured the performance of various recommender algorithms. End-to-end engineering.



Teleport, Inc

• United States
• Technology, Information and Internet
• 1 - 100 Employee

• Data Scientist

  • Jun 2015 - Jul 2015



ASMALLWORLD

• Switzerland
• Travel Arrangements
• 1 - 100 Employee

• Ruby on Rails full-stack developer

  • May 2014 - Feb 2015