Lars Tijhuis

Part of a 5-person engineering team responsible for developing the next version of Post Planner’s product; a social media marketing platform focusing on finding the right content for your audience.

• Transformation of legacy back-end systems (~150K+ LoC, monolithic PHP application with ~6k+ active users) to a microservices architecture using NodeJS and AWS infrastructure.
• Development of a highly scalable service responsible for fetching content from ~340k+ Facebook, Twitter, Instagram and RSS feeds. Improved throughput from 30-180s to 10-20s per feed.
• Development of RESTful API’s; improved response times for high-volume endpoints from 5-15s to 2-3s through storage, query and caching optimizations.
• Implementation of CI approach including Code Quality measures; achieved ~100% code coverage for new services.

Senior consultant for the Software, Infrastructure and Architecture team. My period at KPMG gave me the opportunity to see a high variety of companies while working on their technology related issues and developments.

• Lead on several client engagements 6 months into the job and fast-tracked promotion to senior consultant (2 instead of 3 years).

Some of the impact I made:

• Implemented QA measures in a high-profile software development project (EUR 80M+) for the Dutch government. Performed periodic QA assessments, reported directly to the program manager on project (quality) progress.
• Project lead on the implementation of a social intranet platform at KPMG NL (~5000+ users). The solution was deployed on Azure.
• Defined the data architecture for a telecom provider, determining what data entities to control through master data management processes. Contributed to consolidating 10M+ customer data records to one "truth".

As part of my Master's thesis, I worked on a Dutch spelling correction solution. Using NLP / ML techniques, the solution was able to find spelling mistakes based on the context of a sentence and provide the correct alternative.

• Obtained 1.5M spelling error examples by data mining the Wikipedia revision history.
• The solution was able to find ~50% of the spelling errors based on their context and for 90.6% of spelling errors found, the solution was able to suggest the correct alternative. This was higher than results reported in literature.
• Awarded with an 8 (eq. A).

• Methylium (11/2012 - 6/2013) • One of the lead developers on the Priceline.com and Booking.com apps
• Tam Tam (5/2011 - 11/2012) • Support engineer for Tam Tam's high-profile customers
• Roomplug (7/2011 - 10/2011) • Bachelor's thesis, awarded with an 8.5 (eq. A+) • Implemented an online rent payment portal using PayPal's API's
• Adecs Airinfra (4/2007 - 6/2009) • Rewrote large parts of a Delphi app in C# (e.g. noise pollution models) • Lead developer on Multivariate Data Analysis tool