Shaswar Baban, PhD

Intro

As an experienced professional in the field of Data Science and Architecture, I bring a wealth of knowledge gained from a decade of working in commercial settings. My expertise extends to research and development, where I've spent seven years applying Deep Machine Learning techniques, culminating in the attainment of a PhD.

Through a decade of work experience in Data Science and Innovation, I have a keen understanding of how to create and implement supervised and unsupervised ML models on big data systems, developing dashboards for model monitoring and coaching junior staff on data science. My experience spans a wide range of domains such as pharma, healthcare, retail, legal tech, fraud tech, data quality, social media, and marketing/advertising. I have firsthand experience with startups through London Business School, EntrepreneurFirst, DSV, and King’s20 accelerator programs, and I possess excellent math and statistical analysis skills, gained through intensive PhD courses and international research experience. I am an award-winning academic, with publications in various conferences and journals.

More details on my work experience

Work

I am currently working as a Data Architect and Tech Lead (contract) at NHS Digital in London, UK. In this role, I am responsible for building metadata for the NHS Digital data access environment TRE (Trusted Research Environment) and HDR UK healthcare datasets. I have introduced new metadata entities to the operating model and improved the previous data architecture. I work closely with NHS Trust Hospitals to onboard their datasets and have built many product specification documents for internally/externally collected datasets. I have also proposed a new data flow architecture to automate metadata collection and documentation. Since May 2022, I have been the technical lead and architect of a new metadata product that produces dataset specifications automatically from raw data for all NHS datasets. This has significantly improved the waiting time (10x) for new dataset onboarding and documentation, reduced human resources needed, and improved metadata quality. During this appointment, I have used various programming tools and technologies such as Collibra, MS Power Automate, Python, Pandas, PySpark, Airflow, AWS S3, Archimate, and more.

Previously, I worked as a Lead Data Scientist (contract) at HMRC in London, UK. In this role, I spearheaded the delivery of tens of ML algorithms used for checking the data quality of requests coming from the VAT and Self-Assessment API channels. I have built and tested 25 ML models in production that check the validity and authenticity of metadata sent with the API requests for fraud prevention purposes. I have also developed real-time dashboards and automated weekly reports to track data governance metrics based on the analysis of 100s of gigabytes of data and presented results to stakeholders on a monthly basis. During this time, I have worked with 5 data engineers and data scientists and trained colleagues and junior analysts. I have used various programming tools and technologies such as Python, AWS (EC2, RDS, AES, CW, S3, IAM, etc), Jenkins, Docker, Kala, Elasticsearch, Grafana, PySpark, Dash, HUDI, Airflow, and more.

Before that, I worked as the Lead Data Scientist at Apperio Limited in London, UK. In this role, I developed deep learning (Deep Seq2Seq LSTM) models on AWS SageMaker to predict time-series data such as legal spend and invoice date prediction. I also led the database documentation work and presented to C-level executives and stakeholders of the company. I used various programming tools and technologies such as AWS (SageMaker, QuickSight, Athena, RDS, EC2, S3, IAM), keras, tensorflow, SQL, numpy, scikit-learn, pandas, matplotlib, seaborn, ray, and more.

Additionally, I am a co-founder and CTO of Stan Social Limited in London, UK. I am responsible for the full-stack development of an influencer marketing brand-safety webapp (hellostan.co) called Stan AI. The tool uses deep learning on social media content (text, images from Instagram and Twitter) and generates a background check report on the safety of the influencer's material for promotion purposes. I have used various programming tools and technologies such as AI APIs from Google Cloud services, Python, PostgreSQL, Django, CSS, HTML5, JavaScript, Heroku, and Python Resful API framework.

Finally, I worked as a Senior Data Scientist (contract) at Mirador Analytics where I advised the company on HIPAA compliance and data privacy of a third party's AI & Data Analytics Platform. I led a thorough technical investigation on data security, leakage, and aggregation, which resulted in a deliverable report summarizing possible data breaches and remedial actions.

About

As a highly skilled Data Architect and Tech Lead, I have extensive experience in building metadata for data access environments, proposing new data flow architectures, and automating metadata collection and documentation. At NHS Digital, I spearheaded the technical design and architecture of a new metadata product that has significantly improved the waiting time for new data set onboarding, reduced human resources needed, and improved metadata quality to identify sensitive and personally identifiable leaked patient data. Additionally, at HMRC, I led the delivery of tens of ML algorithms used for checking the data quality of requests and developed an end-to-end data authenticity ML product that analyses over 10 million new requests every month.

I am also the co-founder and CTO of Stan Social Limited, where I have developed a deep learning influencer marketing brand-safety web app using social media content (text and images) that generates a background check report on the safety of the influencer's material for promotional purposes. My experience extends to compliance and data privacy, where I led thorough technical investigations on data security, leakage, and aggregation.

As a certified AWS Solutions Architect, I have worked with a range of tools and technologies such as Databricks, Collibra, Python, Pandas, PySpark, Airflow, AWS S3, Archimate, and more. My expertise in programming, data analysis, and modelling has enabled me to train colleagues and junior analysts, and I have presented results to stakeholders on a monthly basis.

Contact

• Twitter
• Facebook
• Instagram
• GitHub

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';