Nicolai Negru

– Took over an existing Azure Data Platform, analysed pain points and worked on the next stage of its lifecycle by expanding functionality, improving performance and investigating PROD issues. – Maintained and expanded the LLM agent configuration solution, allowing companies to prepare LLMs for their business needs. Designed multi-agent configuration, implemented integration with Kong LLM Gateway, investigated and troubleshooted production issues as part of the prod support.

– Modernised and expanded the pre-trading platform, quantifying the risk of trades that must be cleared through Clearing Houses, as mandated by legislation. Automatically assessing trades in several stages, starting with fast conservative estimates and progressively including more detailed information when more precise margins have to be calculated. – Improved performance of Python code (in one case, reducing execution from 6+ hours to 17 minutes and in another from 4 hours to under 2 minutes).

– Designed and implemented a web application allowing delivery leads to report the progress of their projects and give senior management a clear picture of which projects need assistance to stay on track and within the budget. Integrated the system with 3rd party APIs and systems for authentication and data exchange. – Designed and implemented a PoC synthetic data generator application, which accepted data schema definition and generated large volumes of data of the requested shape/format/volume. – Profiled and optimised the performance and memory footprint of Python applications.

– Maintained and expanded web APIs, processing streams of events from customers navigating the M&S website. Pushing incoming data streams through Databricks pipelines and generating shopping recommendations to be displayed to the customers, based on the analysis of historical shopping patterns.

– Reviewed Pret's existing data pipelines, which the company could not maintain and evolve because of their complexity. – Redesigned data platform to match the company's business needs and technical capabilities, demonstrating to the business and technical groups how new architecture could be used and scaled to accommodate new data feeds. – Implemented the streamlined platform, consuming several streaming and batched data feeds and handing them over to the technical teams for further expansion.

– Maintained and expanded web APIs and data pipelines, consuming trading data to identify traders' behavioural patterns. – Analysing traders' performance in light of historical market conditions to issue behavioural recommendations aiming to increase traders' efficiency. – Onboarded new data feeds, troubleshooted daily data issues, and improved code efficiency and readability.

– Developed and maintained Secure Boot Manager for embedded devices, protecting companies developing IoT software against unauthorised distribution of their products. – Profiled and optimised the performance of Python-based software packages. – Improved CI/CD pipelines, championed TDD and pair programming.

– Worked on several contracts, helping companies establish efficient QA functions, automate tests, and improve/embrace the agile approach in software development: – Lego - JavaScript, AWS – Equal Experts - Kotlin, Python, Kafka, SQL – Gazprom - C#, MS SQL, Python – UnderwriteMe, Java, REST API, HTML5 – Deutsche Bank - Java, REST API, HTML5 – Compare The Market - C#, Python, SauceLabs – Barclays Capital - C#, Python – Microsoft - C#, SQL, XML, Python

– Developing web APIs, automating tests, interviewing, mentoring and managing testing teams, and assessing the feasibility and cost-efficiency of test automation solutions.