Kaz Chichlowski Email & Phone Number

Cracow, Małopolskie, Poland

Supported colleagues in the application of Machine Learning in various settings, including the use of CNNs, decision making models in general and time-series modelling.Developed a SARIMAX-based Statistical Process Control (SPC) for continuous KPI monitoring and assessment for a major US retailer:· Successfully applied to all 1159 stores nationally and transferred to client as a fully working prototype.· Working on the clients GCP with data sourced from GS and forecasts stored in BigQuery.· In-house implementation of ‘auto.arima’ provisioning automated model updates.· Additionally oversaw the contribution of two Senior Data Scientists considering additional ML technologies, specifically LSTMs and XGBoost, intended for inclusion as an ensemble.· Automated summary reporting via googlesheets to prioritise stores for review and model diagnostic reports produced using matplotlib.Proof of Concept / Prototype development for a major US manufacturer for sentiment analysis using WWW to inform marketing and to support industry professionals:• Retrieved the results from applying search terms to Google Trends via a Python API and assisted with applying Facebook Prophet to forecast their future usage.• Overlaid the ranking by geographic region of search terms as returned by Google Trends with a weighting derived from region populations.• Incorporate exogenous data from the National Oceanic and Atmospheric Administration in the U.S. to help anticipate potential near-term demand.• Used the Granger causality test to help identify themes using pairs of the terms received from the Web. Bid support:Engaged with CXOs and SVPs of global and national retailers/manufacturers, delivering presentations on relevant use cases, and participated in exploratory meetings to show how AI and ML may be used to enhance existing operations.

London, United Kingdom

Developed a Data Analytics Environment (DAE) for risk reporting and adhoc querying from multiple data sources:• used for trade data processing, payments data processing, data extract valudation and high risk client identification;• data wrangling via SQLite from sources including Oracle, MS SQL Server, MS Office, Lotus Notes and flat files;• SQL-based API using registered functions developed in-house for content analysis, fuzzy text matching, data summarisation, time/event-line profiling, statistical analysis and scoring;• reporting in Excel/HTML;• written in PERL with optional interface to Python, R and Java.Assumed responsibility for e-comms data and hard disk drive imaging in support of investigations:• used Ruby scripting to produce summary reports on word counts and computer event logs from within NUIX.Helped develop the Case Management System used by DB investigators:• used PowerPoint VBA to implement on-demand case reporting.

London, United Kingdom

• Implemented the Neural Network core of an eFalcon solution for the on-line retailer Shop Direct Solutions scoring credit card usage at the point of purchase.• Developed custom solutions for those clients and helped establish a best practice / standard methodology for the work undertaken by our team.

•Was the point of reference / authority for modelling in our team whose prime objective is to provide data driven customer intelligence to the rest of ClarityBlue.•Used SQL on Sybase, Excel VBA and SPSS Clementine for data analysis and to derive rulesets and scorecards from mixtures of internal, client and bureau data (mainly from Experian & Axciom) for response/accept models.•Lead PoC Implementer for an embedded value-based campaigns optimisation process to run along-side Experian’s Marketswitch for a leading mobile provider.

•Contributed to specialist analysis for health checks targeting fraud for clients including BT, France Telecom and C&W•Provided technical consultancy and support on Data Mining both for and on behalf of Azure Solutions, including a key role in a EURESCOM project [http://www.eurescom.de/public/projects/P1000-series/p1007/default.asp]. Here I implemented a 'champion/challenger' Neural Network based prototype fraud detection system in PERL•Developed a Naïve Bayes Machine (inspired by Paul Graham’s SPAM filter ) as a PoC to both detect fraud and to help in rules management.•Developed the core of a ‘Call Fingerprinting’ module for Azure’s fraud product which became a new feature in a standard build. Also used by Lawful Intervention groups overseas.•Have given presentations on applications of Data Mining to groups including the then Future Technologies Co-ordinating Group, GCHQ.

•Engaged in demonstrating the potential benefits of applying Artificial Neural Networks (ANNs) and other machine learning techniques to the problems of subscription and charge-card fraud, using both SAS and other third party software•Developed a prototype web-based application for monitoring the performance of IP Data Networks including VRML output to map network units such as routers and switches.•Engaged in further research and development activities including the elicitation of rules from Artificial Neural Networks and the prototyping of Guided Knowledge Discovery – a Data Mining paradigm aimed at bridging the gap between domain driven and data driven Data Mining

Phd, Artiticial Neural Networks

Activities and Societies: Captian and President of UEA volleyball team Won UEA sport achievement award, 1992 Thesis title: "Modelling.

B.Eng (Hons), Electronic Engineering

Awarded 1st Class Honours In 1989 I received the Institution Prize from the Institution of Electrical Engineers for distinction.