Full-stack Developer
2021 - PRESENTSelf-employed- Handled the entire lifecycle of original titles, including design, development, release, marketing, post-release analytics, and optimization, mainly using Google Cloud and Firebase.
- Employed evolutionary algorithm-based techniques in Python to generate levels for logic-based puzzle games and optimize key metrics.
- Optimized and automated the localization process for a web-based language learning app, using the Django ORM to enable fast and cost-efficient localization in any language.
- Used Google BigQuery to analyze rich-analytics usage data and conduct statistical experiments, using remote configuration-based functionality changes.
- Developed an experimental Django Channels-based back end that implemented a TensorFlow machine learning model to provide phrase recognition for a language learning app.
- Designed, built, and released a logic puzzle game called Bikit on iOS and Android, using Vue and Capacitor.
Technologies: JavaScript, Python, Vue, Firebase, Google Cloud, Flask, Django, PixiJS, D3.js, SVG, PostgreSQL, CSS, HTML, Nevergrad, SQLAlchemy, TensorFlow, Google Cloud Platform (GCP), TypeScript, Nuxt.js, Game Development, HTML5, Cloud Firestore, Mobile, CapacitorSenior Data Scientist
2018 - 2021Assembly- Acted as project lead for our marketing-attribution solution, which was the team's flagship product and main revenue stream. Built bespoke implementations for various multinational brands and deployed them to production using an AWS-based stack.
- Researched purchase-path-based marketing attribution in depth and implemented all major models in the literature and proprietary models in an R package that allowed an easy application, detailed evaluation, and benchmarking.
- Worked with Google with a beta product they were developing that allowed custom algorithms in Python to be used to optimize programmatic bidding. Later won a marketing award based on the results of a trial with a large hotel brand.
- Researched and applied econometric methods to develop Mixed Media Models and implemented these in R packages, including full documentation and tests.
- Handled the experimental design and implementation of projects to measure the causal impact of marketing strategy changes, including presentation of results to clients and consultation on the next steps.
- Built AWS-based reporting pipelines taking multiple data sources to feed Google Data Studio dashboards for clients in the hotel, luxury fashion, and travel industries.
- Forecasted business metrics using cutting-edge techniques. Developed a productionized product to forecast room bookings for a large hotel client to help inform target setting for their marketing efforts.
- Provided training on SQL-based reporting for the entire organization and helped optimize and troubleshoot existing queries.
- Worked closely with the Cloud Architect to design an improved Airflow and Kubernetes-based stack for the deployment of data products developed by the team.
- Used decision tree models and feature importance techniques to evaluate ranking factors in proprietary search engines of large online retailers to inform on optimization strategies for a large consumer goods brand.
Technologies: Python, R, Jupyter Notebook, Presto DB, Google BigQuery, Amazon Athena, PySpark, Amazon Web Services (AWS), Applied Mathematics, Technical Documentation, Google Analytics, Adobe Analytics, Google Data Studio, Tealium, Apache Airflow, Statistics, Data Visualization, Ggplot2, D3.js, Matplotlib, Pandas, Tidyverse, Dplyr, SQL, Econometrics, A/B Testing, LightGBM, XGBoost, Anomaly Detection, Random Forests, TensorFlow, KubernetesLocalization Manager
2012 - 2018Translate Plus- Founded and grew the Localisation Engineering team to a headcount of six across multiple sites.
- Oversaw the use of computer-assisted translation (CAT), software localization, machine translation, and other technology in translation projects.
- Developed tools and processes to optimize and streamline localization workflows, which were essential for handling increasingly heavier workloads as the company experienced consistent double-figure growth.
- Created and provided technical training courses to ensure the Operations department remained highly skilled as the company grew from 20 to 150 employees across multiple sites.
Technologies: SDL Trados Studio, MemoQ, Computer-assisted Translation (CAT), Machine Translation, Visual Basic .NET (VB.NET), HTML, CSS, Regular Expressions, XML, XSLT, JSON