Building a decentralised cloud, connecting cloud DEPIN projects with Hardware Providers in datacenters around the globe

I lead the team that builds ActiveScale, Quantum's object storage product line, and help our sales teams assist customers in the transition to object storage as the dominant storage paradigm for both warm at cold data.Quantum acquired ActiveScale in March 2020

As a datapath architect, reporting to the VP Chief Software Architecture, I contribute to the Activescale product line in various ways:• Work closely with development teams on object storage and filesystem APIs (Amazon S3 and NFS), the backend data and metadata storage (distributed systems, consensus algorithms, erasure coding), performance measurement and analysis, … to ensure their software maintains our scalability, durability, availability, performance, security, … guarantees • Discuss near-term features and long-term product roadmap with Product Management• Collaborate with my peer architects to create a long-term technical framework that enables successfully building out the product roadmap Apache Hadoop open source contributor• my team contributed dozens of patches to the storage layer (s3a and HDFS) used in Hadoop, Spark, … (especially: s3 committers)• presented HDFS Provided Storage feature (joint work with Microsoft) at the premier Hadoop conference: Dataworks Summit 2018 (Berlin and San Jose) teammate Ewan Higgs became a Hadoop committer based on this workActivescale is an on-premise storage system which customers leverage to scale their storage footprint from hundreds of terabytes to dozens of petabytes in a global namespace with billions of objects/files. I grew with the business as it scaled up from a tech-focused startup, Amplidata, to an acquisition by Western Digital where I work with multiple development teams across 3 continents.Previous roles:• Principal Architect & Product Owner • Senior Product Owner • Senior Software Engineer

I architect the interactions between big data frameworks (Apache Hadoop, Spark, ....) and Amplidata’s Amplistor object storage system and work on some of the internals

A queue is formed when multiple "customers" concurrently require a "service'". This process is omnipresent in everyday life. : • people physically waiting in line (supermarket, post office, GP/hospital, concert, ...)• data packets waiting for transfer in computer networks (telecommunications)• jobs assigned to processors or files to disks (supercomputers, data centers)• supply chain (logistics, manufacturing) Consequently, queueing theory is often considered to be a branch of operations research. It is also a subfield of applied probability as the randomness inherent in the time between customer arrivals and in the time needed to serve a customer is naturally expressed using probability(distributions). The main goal is evaluating the performance of queueing systems (customer waiting time, number of customers in queue, ...).My PhD thesis analytically studies queueing systems with two types of customers (priority and non-priority) where priority customers receive preferential treatment in order to minimize their waiting time. I focus on the difference between allowing a finite or an infinite number of customers to wait in the queue and the resulting impact on the tail of the distributions.Keywords: stochastic modelling, Markov chains, performance evaluation, priority queues, analysis, simulation, (heavy) tails, probability