Engagements
Hi, I’m Albert - engineering lead with experience in building software systems that work.
I help teams to deliver solid software with emphasis on lean and incremental approach, pragmatic programming principles and value delivery without compromising on quality. Here are some details of previous engagements and projects.
ERP AND FULFILMENT – AIR UP
June 2023 - July 2024
Air up (funded in 2018) is a product company that revolutionises the way we drink water by infusing it with scents. I was hired as an engineering lead on Bubbles team and was tasked with helping to migrate heavily customised, untested ERP system (Odoo) into set of maintainable Kotlin/http4k microservices. In the period of 4 months, without any production incidents, the team had successfully migrated EDI integrations of four warehouses from legacy codebase into fully tested, monitored and observable microservices. I then moved onto leading new warehouse migration in United States (Radial to XB) which had been delivered 2 months ahead of schedule and achieved cost saving of $1.2M. Additionally I had been driving the efforts to shape service and tooling infrastructure, testability, and maintainability of new codebase. I promoted best development practices such as use of mono-repository, trunk-based development, continuous delivery with short lead times (<15 mins), shift-left engineering, automatic documentation, strict value typing as well as upskilled other team members versed in Python in writing testable and maintainable Kotlin code.
EQUITY RISK PLATFORM - CITI
November 2021 - June 2023
The Ariane program aims to re-architect and modernise Citi’s Equity Risk platform for live and end-of-day risk across all desks and regions. I was hired as a specialist via JUXT to provide Kotlin and http4k expertise and joined Prism team. The team is tasked with replacing existing platform (based on KDB+, Kotlin and WPF) with more scalable big data solution, that will handle live as well as historical data for all regions. New architecture has been based on Apache stack of Flink, Kafka backed by Ignite and Trino as storage, with http4k services deployed onto internal OpenShift clusters. Due to stability issues and lack of tests in existing systems, for the new services I have mainly focused on implementing logging, monitoring and traceability as well as designing services to be easily testable by use of hexagonal architecture. Prism team has successfully replaced and deployed to production managed hierarchy service (EQMSH), that sourced and served data from several systems integrated via SFTP, and currently is in the process of implementing end-of-day risk system.
TECHNICAL LEADERSHIP - ZÜHLKE
May 2021 - October 2021
Zühlke is a Swiss engineering-focused consultancy with over 50 years of tradition in delivering expertise and digital solutions globally across variety of sectors and industries. I was hired as a technical lead to serve two teams (10 people), taking over development of a part of a customer facing website at Co-op UK.
One of the biggest challenges of this assignment was to form and onboard teams that are fully remote, new to the codebase and distributed in United Kingdom, Portugal and Serbia. As team had one week of hand-over, in order to aid knowledge transfer from leaving delivery team, I have organised and recorded sessions on different aspects of development and infrastructure. To further empower the teams and enhance onboarding I have organised and led daily mobbing sessions, which then continued with rotating leader. This ensured collective understanding of codebase, resolved configuration issues and shared best practices. I have setup daily and weekly ceremonies, helped define ways of working, reduced onboarding friction and provided all necessary tools and access for the team members to feel empowered in daily tasks. Currently teams are remotely pairing using Tuple and working on weekly Sprint cadence with a view to move to Kanban. As a floating team member, I am also involved daily in coding and working on stories. Technology stack includes Python, Django, PostgreSQL, Amazon AWS infrastructure (Lambda and EKS) as a backend as well as JavaScript and Vue.js as frontend.
As a lead I am also heavily involved in technical leadership meetings, which aim at identifying technical debt, coordinating projects and improvements across all teams. I have led process that aimed to identify and improve product delivery in balance with developer experience, identify maintenance and decision debt and present the plan to Product Council.
TECHNOLOGY ASSESSMENT FOR DELIVERY TRANSFORMATION - THAMES WATER
November 2020 - December 2020
Rapid Delivery Transformation programme was a Thames Water effort to improve efficiency of delivery teams. I was engaged on temporary basis as tech lead working with Product Working Discovery team along CTO and leadership. Team was helping to identify process and technology improvements across 32 teams and is 6 technology groups. Collaborated on and defined framework, which served as a baseline for assessing teams and organisation through product, delivery, technology and people lenses. Through the series of guided workshops, process delivery mapping sessions, surveys, individual interviews and artefacts, assessed maturity of delivery, DevOps adoption, testability, monitoring, stability and best practices across technology groups. Collaborated on and delivered set of principles that aimed to accelerate design and development, and deliver valuable solutions rapidly. Identified and presented improvement to delivery process in several groups with varying delivery models and agile maturity levels.
STRONG CUSTOMER AUTHENTICATION AND APPLICATION SECURITY - BÓ (NATWEST)
April 2018 - October 2020
Bó was a Natwest initiative to create current account companion app that helps customers to manage money better. In early stages of the project I was responsible for building backbone for micro-services and setting up best practices across the stack, such as build and deployment pipeline, app testability (contract, acceptance, external and integration), monitoring, structured logging and automatic documentation. I have been involved in building customer onboarding journey where I have successful integrated with several vendors and systems, including Equifax, Hooyu, Vault, LDAP, Ping Access & Federate and IdentityX as well as implemented equivalent vendor fakes.
Later I moved on to leading IDAM team building PSD2 compliant Strong Customer Authentication service as well as implementing and overlooking service-to-service and app security across entire bank. Team implemented iOS and Android specific hardware-backed cryptography module (utilising components such as Secure Element, StrongBox and TEE) to provide security for keys, signing and encryption as well as utilised Diffie-Hellman key exchange for session encryption. Backend services had been written in Kotlin and utilised HashiCorp’s Vault for cryptographic primitives. As a part of the solution team also implemented LDAP-based employee authentication system and distributed authorisation policy protecting endpoints within service mesh. Systems had been deployed using Helm to AWS EKS. I have also organised threat modelling sessions, security assessments with external party as well as system migration and rollout strategy to ensure project was delivered meeting bank-level security requirements.
AEI & FATCA REPORTING PLATFORM - CREDIT SUISSE
May 2017 – April 2018
Fondue team was a core part of Credit Suisse Big Data initiative. The main objective for the team is to bullet-trace and deliver data platform, which provides access to APAC’s private and investment banking data, with a view to expand to other regions. Additionally, to drive out the solution, team is delivering set of reports and system feeds to cater for upcoming wave of AEI (Automatic Exchange of Information) and FATCA regulations. As a strategic project, the team have all necessary flexibility and technical freedom in making decisions that shape and improve the platform. Fondue platform is a set of rest applications, React Web UI and Apache Spark jobs running on top of internally hosted Cloudera’s Hadoop distribution. Data was made available to other teams via HDFS and through Apache Impala, which allowed building basis for trade recommendation engine.
FOUNDER - IGNECO
July 2016 – April 2017
Building out technology foundation in early stage start-up. Waggol is a mobile first activity marketplace aggregating outdoor activities. Deliverables included hands-on proof of concept (JavaScript, Node.js, Heroku, MongoDB, React/React Native), further technology outline and assessment, competitor analysis, costing, delivery estimation and investor’s deck preparation.
I founded of a small team building securities lending platform (Secle) for prime brokerage. Project involved implementing foundation services for trade capture/flow for securities collateral trades, market data feeds, billing, PnL and exposure calculations. I have also implemented a generic user management and application permissioning service (Autherio) based on JSON Web Tokens. Both projects were written in Java 8, PostgreSQL, built with CircleCI, packaged with Docker and deployed to Amazon Web Services.
BIG DATA PLATFORM - ELSEVIER
February 2016 – June 2016
Working on Link Expansion project, which is a part of Elsevier’s Big Data department. Elsevier is academic publishing company dealing with large volumes of journal metadata. The main objective of the Link Expansion project is to become company-wide source of document link by aggregating different sources and APIs. As a senior developer I was involved in building several source aggregators and crawlers (>100 million links). Services were built using Scala, Apache Spark and deployed on AWS Elastic Map Reduce (EMR) and DataPipeline infrastructure. I was also responsible for setting up build pipelines and deployment of services using Jenkins, JobBuilder and Terraform.
TRADING SYSTEMS - THEOREM TECHNOLOGY
June 2014 – November 2015
Leading team of five in custodial and trading department at Solo Capital Partners and then Theorem Technology. Theorem Technology is a FinTech software consultancy building scalable, bespoke trading platforms fully on Amazon Web Services (AWS) infrastructure. As a lead I was heavily involved in orginising team around building order management, custodial and Know Your Customer (KYC) systems as well as various supporting services. Main technology stack included Java 8, Clojure, PostgreSQL, RabbitMQ, TinkerPop and AWS (EC2, S3, ELB, CloudFormation, RDS, SQS, SES). I was also involved in simplifying delivery practices by creating command line infrastructure tools and setting up automated pipelines. All systems allowed for automatic, horizontal scaling and had continuous deployment pipeline to test environment with some systems being continuously deployed to production.
PRICING POLICY AND FUNDING NAME GENERATION - DEUTSCHE BANK
July 2013 – June 2014
The platform aimed at reducing and optimising collateral cost and exposure across the bank. System consisted of three main components: Policy (pricing policy system), Fuse (funding name generation service) and Gretl (collateral agreement capture). Initially I helped developing Fuse (Scala, MySql), system providing traders and systems with correct, bank-wide funding name based on counterparty and agreement in place. Consequently I have moved over to work on Policy (Scala, Clojure) which defined set of pricing policies dependent on organizational and regulatory constraints. As a senior developer I was responsible for implementing features and fixes across and interchangeably between Fuse and Policy systems, contribute to design, participate in planning session and run retrospectives.
PRIORITY MOMENTS - O2 (TELEFONICA)
February 2013 – July 2013
The Priority Moments is a service that provides users with means to redeem location-specific voucher codes and offers for listed merchants from mobile device. The system consisted of four main components: Global Offer Service, Offer Manager, Global Ingest and Reporting Service. All services communicated via HTTP/REST interfaces and are deployed and managed on Amazon Web Services. The Global Offer Service (Scala, Unfiltered and MongoDB) is responsible for providing REST services to manage merchants and offers as well as execute batch publishing jobs to activate new offers. Offer Manager (Groovy, Grails and MongoDB) acts as an administration interface for sales and support teams. Global Ingest (Scala, ActiveMQ, MongoDB) consumes and processes application events, which are then aggregated by Reporting Service (Scala, Unfiltered, MongoDB) and fed back to Offer Manager. As a senior developer I was responsible for implementing new features across all services, contribute to design and participate in planning session. I have also contributed in establishing paring culture within the team which helped to decrease friction in several areas.
BROADBAND PROVISIONING SERVICE - SKY
May 2011 – February 2013
The team was responsible for developing Halo and FTTC systems which purpose is to automate provisioning workflow for voice and broadband (DSL and fibreoptic) services. The systems are implemented using core Java, Tomcat and communicate with dowstream systems via ActiveMQ and REST. The team is fully agile/XP with high emphasis on quality of delivery achieved through pair-programming and several layers of testing covering acceptance, contract, integration and performance. The tests had been written in Concordion and are currently being ported to more refactorable open source framework called Yatspec.
I was significantly involved in design and implementation of order event feed service, which main goals were to provide strangler pattern over legacy schema to allow refactoring as well as to expose order processing data to support and reporting systems. I was also driving refactoring efforts towards simplifying message handlers by introducing more concise rule matching engine and separating provisioning of different products into stand-alone systems.
I have organised hands-on brown bag session “Clojure Testdrive”, which presented high level overview of Clojure programming language and gave opportunity to use it to solve several example problems. I have also proposed idea and participated in a hackday, for which my pairing partner and I have implemented mobile broadband speed checker using node.js and PhoneGap.
COUNTERPARTY EXPOSURE MANAGEMENT PLATFORM - RBS
May 2011 – November 2011
Cremant is a credit risk calculation platform that provides global CEM desk with exposures and associated measures (e.g. sensitivity, PnL, CVA, FVA, DVA, WWR) for specified counterparties. The system consists of three main tiers: frontend C# client, SOA services and DataSynapse calculation grid, all backed by Sybase database. The services have been implemented using Java clustered with Coherence, internally communicating over RMI and exposed over JMX, WSDL and SOAP.
As a lead developer I worked closely with quantitative and business analysts and was responsible for delivery of two core projects: collateral calculation modelling, which involved implementing quantitative models for calculating collateral (e.g. local/hard currency collateral, independent amounts), and credit curve explosion, which aim was to generate different pricing curves based on counterparty rating and liquidity factors. Projects required changes to calculation engine as well as exposing relevant operations and data to the client over web services and to reporting module over JMS.
OTC PRODUCT LIFECYCLE MANAGEMENT PLATFORM - UBS
May 2010 – May 2011
Klondike is a product lifecycle management platform, which core functionality is processing securitised and OTC product lifecycle events such as product creation, amendments, trades, barrier breaches, corporate actions and expirations and dispatching them to front and back office systems. The system is also acts as a source for product and trade information and is responsible for settlement valuation for expirations, early exercises, batch processing of corporate action adjustments as well as syndication of financial reports and barrier breach events to external parties. In terms of technology Klondike has been implemented using Tomcat with Lucene/Compass searching, Webwork, XWork and FreeMarker, backed by Oracle database. The team was fully agile/XP and put particular emphasis on delivering high quality software. Daily pair programming sessions, use of test driven development and maintaining test coverage with over 12,000 unit, integration and acceptance tests allowed team to produce to the highest standards.
I was working on building Barrier Breach Syndication Service, which captures barrier breach events received from monitoring system and syndicates them to third parties using various transport protocols such as S/FTP, HTTP/S, SCP or email. The service has been deployed on Tomcat server and is currently syndicating events to 6 external vendors.
I was responsible for delivery of core components for Mass Issuance project. The main goal of the project was to allow mass issuing and straight trough processing of leveraged (Warrants, Mini-Futures) and yield enhancement (Discount Certificates) products across German and Swiss exchanges. The work included booking product to risk and settlement systems, implementing expiration, early exercise and corporate action calculations as well as optimising batch processing. After successful delivery UBS gained 1% market share on Swiss and German exchanges within the first month of trading.
ADVERT RETARGETING PLATFORM - VIBRANT MEDIA
November 2009 – May 2010
Established in 2000, Vibrant Media is a contextual media pioneer and leading provider of contextual video advertising. I was working on extending In-Text platform, that serves contextually relevant advertisements to over 100 million user world wide. The system was developed using JavaSE making use of NIO and concurrency framework and deployed onto 200 Tomcat servers clustered in three geo-locations, backed by MicrosoftSQL/MySQL databases. I was responsible for implementing user retargeting module, which was used in behavioural profile analysis to improve revenue by retargeting delivery of campaigns at users with specific profile characteristics. Due to high load and aggressive time constraints on delivery of inventory, the project had been developed with particular emphasis on performance and thoroughly tested with unit, regression and performance test.
MOBILE SPREAD BETTING PLATFORM - IG INDEX
January 2008 – November 2009
IG Index, founded in 1974, was the first financial spread betting company in Britain, and is today the leading operation of its type in the world. I was leading a team of 3 working on redesigning of dealing platform for mobile devices. The system consisted of three tiers: backend service, which interacted with core dealing services; standalone middleware server, which provided communication bridge for mobile devices, encryption and live price streaming; and frontend JavaME mobile clients. The backend service was implemented with EJBs, Spring and Hibernate backed by Oracle database and deployed onto Weblogic cluster. Middleware server was based on Spring, Jetty, Java Concurrency, JMX and integrated with proprietary price streaming server over JNI. I was also responsible for designing and implementing framework that will allow porting of existing mobile dealing application to NTT DoCoMo Java (Doja). The main tasks were implementing network protocol stack using sockets for client and server as well as developing UI bridge. The application has been successfully introduced as a part of FXOnline launch and covered more than 15% of clients.
I have successfully introduced elements of agile/XP within the team, which improved communication, code quality and delivery time. In order to increase code quality, we have followed test driven development (TDD), conducted frequent code reviews and pairing sessions. Additionally, I have implemented a set of integration and regression test domain specific language (DSL). Tests were executed during nightly regression build on continuous integration server and helped detecting several, critical issues prior release to production.
TEXT RELAY SERVICE - DSPG TELECOM
July 2004 – December 2007
Established in 1988, DSPG is an ISO 9001 registered company specialising in Digital Signal Processing (DSP) and telecommunication equipment. The main projects I worked on were Text Relay Service (TRS-1) and TexBox-ISDN. The TRS-1 is a media relay service and was developed for BT TypeTalk to connect 150 operators providing media translation services for hearing or speech impaired clients. I was a responsible for developing server and clients that bridged different communication media, such as email, FAX, SMS, internet and mobile chat. The backend was designed using Spring, Hibernate backed by Jetty and MySQL database, communicating with ISDN stack via proprietary TCP/IP protocol. I have successfully integrated the service with third party SMS provider (HSL SMS), FAX and IMAP email services. As a frontend to the system I have implemented three clients: JavaSE/Swing based operator’s terminal, JavaSE applet for internet chat and JavaME mobile client. Clients shared common communication protocol which was optimised for mobile communication. The TRS-1 systems are currently used by TypeTalk in UK and relay services in Spain, New Zealand and US. The second project I worked on was TexBox-ISDN, which involved implementing JavaSE/Swing based operator’s terminal as well as integrating C++ VOIP library for voice streaming. The TexBox ISDN systems are used by number of organizations including Transport for London, British Deaf Association, Westminster City Council, NHS Direct and National Rail Enquiries.