Boon Ping Lim Email and Phone Number

Building data product & data science team in Telco business

Combining Machine Learning and Graph Analytics with Operation Research for problems related to- Voyage planning/optimization- Fleet optimization- Graph networks of journeys- Trade flows of commodities- Performance optimization with vessel specific sensor data- Hierarchical time series predictions

- Building graph analytics pipelines. - Graph-based deep learning (GCN / tensorflow), - In-memory graph (networkx, GraphFrame, cuGraphs),- Graph DB (JanusGraph/ HBASE/ elasticsearch) on Hadoop/Spark- Application domains: knowledge graph, unique user identification, fraud detection, trust & safety, credit scoring, search & QA, recommendation, product category.

• Involve in the design and development of stellargraph.io, a graph-based Investigative Analytics Platform.• In-charge of designing and developing Entity Resolution framework at scale, for ~1 billion nodes record linkage and deduplication.• In-charge of designing and developing community detection and new pattern discovery approaches for fraud and criminal detections, focusing on money laundering, terrorism financing, organised crime and tax evasion.• In-depth knowledge in entity resolution technique (using deep learning, meta-blocking approach), graph-analytics (using networkx, graph-tool, GraphFrame), distributed computing framework (Apache Spark, Flink), machine learning libraries (MLlib, scikit-learn), deep-learning framework (tensorflow, GCN, node2vec).• In-depth knowledge in continuous integration continuous delivery tools: github, Maven, docker, travis CI• Collaborating with external customers such as Australian Taxation Office (ATO), Australian Transaction Reports and Analysis Center (AUSTRAC), Australian Federal Police (AFP) to understand problems and their data, and communicate results and insights to audiences with differing technical backgrounds.• Collaborating with research, engineering and business teams across Data61 to ensure that project goals and Data61’s goals are achieved.• Providing technical direction, taking ownership of projects, presenting ideas and fostering creativity in others.

• Focus on energy management in smart buildings, specifically on heating, ventilation and air-conditioning (HVAC) operations control and occupant activities scheduling, with an aim to reduce energy consumption in commercial and educational buildings.• 3 years hands-on experience in mathematical modeling and optimization techniques. In-depth knowledge in linear programming (LP), mixed-integer programming (MIP), large neighbourhood search (LNS) using Python and optimization tools such as Gurobi, CPLEX, SCIP, and IPOPT.• Good knowledge of machine learning techniques such as linear regression, logistic regression, support vector machine, k-means clustering, neural network and recommendation system, and data analytic tools such as R, SciPy.Research Highlights/Outputs:• Designed and developed: algorithms and mathematical models to solve joint HVAC control and occupancy scheduling problems. mathematical programming models to address the challenges of combining discrete scheduling problem and non-convex non-linear HVAC control problem. machine learning techniques to improve HVAC and building models using large-scale building datasets. a large neighbourhood search (LNS) algorithm to handle large-scale models. an online scheduling model to cope with impromptu job requests and cancellations. a robust optimization model to dynamically control occupied thermal comfort bounds based on outdoor temperature and user preference• Performed model validation using building energy simulation software Energy+ and Simergy.• Others: Developed a mixed-integer non-linear (MINLP) model for HVAC control. Explored hybrid discrete-continuous model using constraint programming (CP) model.• Filed an international PCT patent entitled “Controlling operation of energy-consuming devices” on the core techniques and the system of the research work above on Jul’15. • Published and presented 3 conference papers in top-class conferences in the area including AAAI.

• Conduct R&D in Cloud Computing project, which was funded under the Malaysia’s 10th Plan (2010-2015). • Performed R&D on resource provisioning techniques on MIMOS Cloud Computing platform.• Designed a network utilization prediction algorithm to estimate the various network resources utilization (such as latency, bandwidth, memory and processing power) for delay-sensitive cloud applications.• Implemented the network utilization prediction algorithm using C/C++ on Kernal-based Virtual Machines (KVM) and Ubuntu Linux.

• Focused on R&D of audio video (AV) communication and its applications. • Led a team and developed, in close collaboration with Panasonic worldwide R&D labs, various R&D projects in multiparty video conferencing, AV streaming, AV transcoding, home media network.• 6 years of experience on ARM-based embedded system development with C programming. In-depth knowledge in P2P communication and optimal streaming path selection.Responsibilities:• Pioneer member of Panasonic Kuala Lumpur Labs. Oversaw a team of engineers specialized in AV communication protocols.• Drove R&D activities and software prototype development based on SDLC/CMMi processes.• Articulated, compared, and implemented solutions and alternative approaches based upon project management principles and R&D targets.• Main interface person to manage and negotiate with customers on project requirements, project deployment schedule, system design, functional & performance test scope.• Identified functional gaps and translated business problems and requirements into technical design.• Regularly represented the team to Japan and IETF meeting for business promotion, project meetings, presentations, demonstrations and integration activities.Accomplishments:• Delivered application layer multicast feature into Panasonic KX-VC series, from research idea, to PC-based prototype, and finally to production level in 3 years. • Designed and developed optimal routing path, fast routing convergence and bandwidth fair algorithm for multi-path HD audio-video streaming.• Design and developed high-speed kernel level packet replication to perform application layer multicast.• Design and development of TCP/UDPIP based networking software for large-scale distributed systems over Internet and home network.• In-depth knowledge in path planning and optimization algorithms, real-time streaming system design and performance optimization.• Knowledge in embedded systems (Intel IXP platform, Panasonic Uniphier platform).