Abdul Hakim Norazman

Scaling and hardening a Java/Spring microservices platform used for large fund-administration reporting workflows.

• Eliminated an N+1 database lookup in a high-volume statement-generation service, reducing allocation-type lookups from 103 to 2 per request and removing about 220,000 production database round-trips per business day.
• Redesigned a returns-calculation data fetch to make one lean, deduplicated upstream call per fund, making the median fetch 13.9 times faster and reducing the worst production case from 31 minutes to 10 seconds.
• Fixed deterministic out-of-memory failures on the largest client funds by right-sizing data fetches; UAT showed a previously failing fund completing and another running about 36 times faster with 15-57 times less allocated and promoted memory.
• Deduplicated redundant upstream calls across parallel calculation branches, reducing calls per run by 64% and avoiding about 6,900 production calls per month.
• Restructured a workflow dependency graph so one task failure invalidates two downstream tasks instead of eight, a 75% reduction validated in UAT, preventing unrelated investor-report work from waiting on the failure.
• Diagnosed JVM and container mis-sizing across four services, correcting a 127-CPU JVM view on 2-4-core containers, a profile-selection heap defect and manual production memory overrides through configuration as code.
• Eliminated a recurring production database failure caused by temporary-table name collisions under concurrency, reducing 2,285 incidents across 84% of business days to zero after release, confirmed in two independent log sources.
• Fixed connection leaks behind a production database-pool exhaustion incident and added pool-usage gauges; a load test drove the pool to its 20-connection limit, completed all 204 requests and returned usage to zero within about 30 seconds ahead of production release.
• Modernised a financial-statements Bulk Override Tool by upgrading Java, Spring Boot and Gradle.
• Worked across engineering and financial-domain boundaries to diagnose bottlenecks, clarify requirements and deliver maintainable production changes.

Built research tooling used to model, review and report long-term market assumptions.

• Developed a Java and React platform supporting Long-Term Capital Market Assumptions.
• Delivered an Asset Volatility module with backend calculations, controlled manual overrides and report integration.
• Upgraded Spring Boot versions across services using OpenRewrite and GitHub Copilot to improve maintainability.
• Tested and documented REST APIs with Swagger to support reliable data flow between research tools.
• Worked with research and engineering stakeholders to translate analytical workflows into production software.

Developed and operated cross-asset trading systems, primarily as a Java server-side engineer.

• Migrated millions of reference and instrument records from legacy Sybase databases to internal platforms using Kafka, REST and GraphQL, with SQL-based reconciliation.
• Delivered instrument-specific data to Portware to improve ETF order releases and reduce trading latency.
• Enabled warrant-note orders in Asia to flow into Portware, reducing manual trader intervention and operational risk.
• Strengthened secure connectivity to external services including Bloomberg, OTCX and Symphony.
• Resolved urgent testing and production incidents across equities, fixed income, OTC, money markets and cash-management order routing.
• Acted as a global release manager, coordinating production delivery, support communications, system monitoring and trader-facing release notes.
• Moved order-grouping eligibility logic from the user interface into backend services ahead of a UI modernisation.
• Built relationships with traders and portfolio managers through one-to-one feedback sessions, workflow study and order-routing subject-matter expertise.

Part-time technical role supporting a digitally focused makerspace alongside my engineering degree.

• Helped manufacture more than 10,000 3D-printed face shields for NHS frontline staff during the COVID-19 pandemic.
• Delivered general inductions on the safe use of 3D scanners and printers, VR/AR equipment, CNC tools and electronics.
• Created and taught 90-minute interactive workshops in Inkscape vector-graphics design for students and staff.
• Worked with the School of Islamic and Middle Eastern Studies to research, design and manufacture an accurate, scalable astrolabe.
• Acted as a subject-matter expert on digital-manufacturing equipment including laser cutters.
• Contributed to the team recognised for Outstanding Library Team at the 2020 Times Higher Education Awards.

• Thesis: The Development of a Machine Learning Computer-Aided Design Filter Using Computational Geometry.
• Completed the thesis in partnership with Theorem Solutions and Ford Motor Company, using C++ and Python.
• Combined mechanical and electrical engineering with software, mathematical modelling, project management and manufacturing.
• Worked part-time as a Makerspace Technician in the University of Edinburgh's uCreate Studio.

Software & digital systems

C programming, C++, Python, MATLAB, R, Verilog and digital-systems laboratories

Mathematics & computation

Matrix analysis, differential equations, Laplace transforms, probability, statistics and computational geometry

Engineering & delivery

Mechanical engineering projects, supply chain management, industrial management and engineering project management

Led the stability, braking and suspension programme for HYPED's third-generation pod while representing the UK at the SpaceX Hyperloop Pod Competition.

• Led three technical sub-teams totalling around 25 people across stability, braking and suspension.
• Managed a technical budget of up to GBP 25,000.
• Represented the UK as a finalist at the 2019 SpaceX Hyperloop Pod Competition IV, where HYPED finished in the top ten.
• Contributed to HYPED's wider competition record, including sixth place at the 2018 SpaceX Hyperloop Pod Competition.
• Delivered a sold-out Institution of Mechanical Engineers lecture on Hyperloop technology and HYPED to an audience of around 100 people.
• Presented technical work to external audiences across the country.