Si Kai Lee Email & Phone Number

San Francisco, Ca, Us

San Francisco, Ca, Us

Worked with Mladen Kolar and Panos Toulis.

Tübingen Area, Germany

Worked with Mijung Park and Krikamol Muandet on analysing the effect of privatised propensity scores on downstream causal inference tasks via inverse probability weighting.

Tokyo, Japan

Member of the Approximate Bayesian Inference team at the RIKEN Center for Advanced Intelligence Project headed by Mohammed Emtiyaz Khan. Worked on continual learning and learning representative subsets for correlated latent variable models, specifically Gaussian processes.

Oxford, United Kingdom

Member of the Reinforcement Learning (RL) team led by Joao Messias. Learned the basics of RL and Imitation Learning, and surveyed state-of-art policy learning methods. Baselined Ho and Ermon’s Generative Adversarial Imitation Learning for a proof-of-concept task and implemented a variant of the variational autoencoder in Wang et al’s Robust Imitation of Diverse Behaviors. Wrote a brief tutorial on natural gradients and how they are used in policy gradient methods.

San Francisco Bay Area

Part of the analytics team working on converting user data into actionable insights. Built data pipelines and dashboards on Periscope and local servers and aided in the migration to Amazon Redshift. Applied machine learning algorithms such as Naives Bayes SVM, TF-IDF and Topic Models in conjunction with packages such as Scikit-Learn, SpaCy and Vowpal Wabbit to predict user sentiment and behaviour. Gave a presentation on Natural Language Processing representations and machine learning techniques.

Singapore

Developed tools such as data processors, parsers and vectorisers, implemented clustering algorithms and carried out other tasks relevant to the research of Associate Professor Wei Lu in statistical natural language processing.

Singapore

Identified and documented problems experienced in the Boeing 767 Freighter Conversion Program and recommended appropriate solutions for Boeing’s approval. Communicated with both technicians and Boeing personnel to ensure that the issues were well-understood and explained. Provided innovative solutions that reduced cost to various structural and wiring problems.

Singapore

Tasked to find suitable connectors for a new PCB control panel for future wind turbines. Communicated with distributors to obtain samples and did an in-depth cost- benefit analysis of the different connectors. Demonstrated problem-solving skills by devising different wiring configuration for different connectors. Prepared a report on the viability of different connectors for different parts of the panel.

Doctor Of Philosophy - Phd, Statistics

Master Of Science (Ms), Computer Science (Machine Learning)

Aeronautics And Astronautics, First Class Honours

Activities and Societies: Climbing Team