At BrainboxBI, using AI and Data Science expertise, we offer AI-as-a-service, BI-as-a-service, and Platform-as-a-service. What We Offer:AI-as-a-Service: Utilise advanced algorithms to extract actionable insights from your data, driving innovation and improving decision-making across your organisation.BI-as-a-Service: Our Business Intelligence services empower users to create impactful visualisations and dashboards, facilitating informed business decisions.Platform-as-a-Service: Optimize your operations with our comprehensive platform, designed to support planning, analysis, and optimisation across various functional areas.We believe in putting the power of data into the hands of business teams, with analytics teams as their enablers. Join us in transforming your data into actionable insights and driving your business forward.

Key End-Client projects undertaken:Electric Power Research Institute• Development of Named Entity Recognition model using spaCy • Design and implementation of fully productionised model involving web scrapping (Python)Parkinson’s UK • Unsupervised learning of Parkinson’s disease cohort: K-means and PCA clustering • Development of disease severity score• Time-series analysis of Parkinsons’s disease progression• Dynamic visualization of patient’s “journey” across a “disease map” • Predicting patient’s rate of disease progression using AutoML Touchnote • Churn prediction modelling: Logistic Regression for EDA and XGboost for model production (R)• Fully automated model production: AWS-instance-run, cron-based weekly R process to write probability of churn scores for currently active users onto company live databseParkingEye • Open Street Map API gathering of geographical features data: bounding-box methodology development• Development of model for an estimation of ROI given geographical information: GLM framework, random forest feature importance, xgboost final choice model (average error of 39%)• Model integration with fully functioning RShiny appNews UK • Time series traffic modelling: o GCP-based full model production (fully automated)o Regression & autoregressiono Traffic Index metric design and development• Long-term customer engagement modelling:o Evergreen Index metric design and development.o Documentation-based model production (semi-automated)o Exponential and max-transformed data modelling.• Sentiment analysis via Google API: assessing impact of article sentiment on customer engagement• A/B testing of The Times Freemium paywall-free model:o Design of A/B testing framework (randomization)o Determination of monitoring metrics and obtaining them using SQL via GCP platformo Statistical analysis of data in Ro Data visualization in Google sheet• LSTM propensity to subscribe modelling• Comment toxicity modelling

• Researching and building customer churn predictive models (R & Python).• Building targeted and individualised reports and dashboards (R, looker).• Providing insights into data and sales methods (SQL).• Broadening understanding of customer behaviour with data (R).• Project management via Jira.• A/B tests design.

Key duties:• Design of Monte Carlo simulation tool for booking curve optimization (Python):- Modelling price elasticity and market share with an aim of maximizing revenue• Design of calculus-based solution for booking curve recommendations based on price-driven elasticity of demand equation (R)• Delivery of enterprise analytics (R):- Studying the effect of customer cancellations on bookings and revenues- Assisting stakeholders involved in predictive modelling and forecasting- Studying the effect of airline flight capacity changes on company’s profits• Developing R package and Python module containing tools for revenue management.• Design of cloud-based solutions for data visualization.• Using SQL to query data from company’s database.

Key duties: • Designing algorithmic pipelines for reproducible and rapid analyses of NGS data in a Unix shell.• Developing R package containing a library of tools for bioinformatics analyses.• Data analytics of high throughput medical and biological data using R and Python programming languages.• Accessing gene ontology and epigenetic annotation data using R-based BioMart tool. • Solving computational problems and exploring new methods for bioinformatics investigations: (1) overcoming the limitations of high gene drop-out rate in single cell RNAseq, (2) designing machine learning algorithm for biological sample classification• Statistical analysis of data: (1) parametric tests (t-test, simple regression, multiple regression), (2) Non-parametric method (Fisher’s exact test, rank based approaches), (3) Bayesian approaches • Mathematical modelling of data: fitting linear and higher order polynomial models using least squares.• Using machine learning for sample classification and dimensionality reduction: application of support vector machine, PCA and t-SNE.• Providing support to experimentalists: (1) writing documentation associated with the developed R package for easy use by the end user, (2) oral explanation of data analysis.

Key duties:• Analysing blood samples: automated and manual cell counts testing, measuring concentrations of biochemical compounds, and supervised results interpretation.• Sample receipt and booking into hospital computer system.• Liaising with theatre support stuff for tissue transport.