Motivated Individual with keen interest in applied machine learning and data wrangling. I am not afraid of data janitorial work. Experience with end to end data product development and maintenance. Experience in translating academic research into production. VIM user and mechanical keyboard enthusiast.
Summary
Experiences
Working service type recommendation for cross sell and contextual messaging
Worked closely with Revenue Management's Ancillary line of business to oversee the creation of datamarts, dashboards and building data products such as dynamic pricing of baggage, seats and food items. Managing team of 5: 1 Data Analytst and 4 Data Scientists
- Dynamic Pricing, perishable inventory based on departure date, using trip, flight and passenger attributes
- Price Elasticity Experimentation at route origin-destination level. Taught business analysts how to use dashboards to inform pricing decisions
- Personalisation (Take-Up Rate Models) and deployment of FastAPI based proxy on GCP CI-CD on GAE and featurestore using FireStore
- Cost reduction: Inflight perishable inventory, ie. hot food and help with demand planning.
Set up interface for production and test environments to use data science models
- Set up rulesets / ML guided predictors within Pricing Engine (Ancillary Pricing Optimization APO) for Seats / Meals / Baggage
- Set up Price Exploration Dashboards
- Factor Based ML Models
- A | B testing using Navitaire’s Analytics Module
- Demand Planning (F&B: Fresh and Perishable goods)
- Propensity Model for EDMs and Recommendation Engine
Specialized in E-commerce (Groceries and Food Delivery) personalisation
- Recommendations engine model development, using (i) collaborative filtering (He. et al), (i) associations rule mining and combination DNN using embedding based features with linear regression, wide and deep.
- Set up recommendations service using Flask API, OpenAPI, pytest, Bug tracking (bugsnag)
- APM and Tracing (Datadog) and A|B testing framework PlanOut (from facebook)
- CI/CD Deployment to K8S cluster (AWS), responsible to writing and maintaining helm charts and terraform for machine learning projects based on sklearn, heuristic models, and deep learning models using tensorflow serving
Acted as CTO and Co-founder for company. Acquired in 2019 by DollarsandSense.sg (Number 2 largest financial literacy site in Singapore)
- FE and BE web development (node.js ES6), µservice API (Python)
- REST APIs integration with business partners eg. DollarsandSense.sg
Publications
Below is list of my past Publications
RiboTagger -
Xie, C, Goi, CL, Huson, DH, Little, PF, Williams, RB (2016). RiboTagger: fast and unbiased 16S/18S profiling using whole community shotgun metagenomic or metatranscriptome surveys. BMC Bioinformatics, 17, Suppl 19:508.
TF Motif enrichment -
Goi, CL, Little, P., & Xie, C. (2013). Cell-type and transcription factor specific enrichment of transcriptional cofactor motifs in ENCODE ChIP-seq data. BMC Genomics, 14(Suppl 5), S2. http://doi.org/10.1186/1471-2164-14-S5-S2
Deepore -
Ang, ML, Raja A, Goi, CL, Deepore: Sequence-2-Sequence machine translation model of pore based DNA sequencer electrical signals to DNA bases
MetamapsDB -
Goi CL, MetamapsDB: Network Analysis of Metabolic Pathways and Gene Based Assemblies in Metagenomics R package on CRAN (2017)
Projects
Below is list of my past Projects
Predicting Mortality: Tidal Volume -
Mortality Prediction using EMR and log data from respirators during in-patient stay (Tidal Volume)
Best Paper -
Presented at The International Conference on Bioinformatics 2013 (Taicang, Best Paper Award)