Saatviga Sudhahar Shaseevan
Senior Machine Learning Scientist at Healx- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
English Full professional proficiency
-
Tamil Native or bilingual proficiency
-
Sinhalese Limited working proficiency
Topline Score
Bio
Andrea Pierleoni
I worked with Saatviga for two years at Healx on advancing deep learning methods applied to biomedical knowledge graphs to discover new drugs . Saatviga has a deep understanding of deep learning practices, she delved on a completely new research topic, got a good grasp of the literature around it and delivered multiple solutions thoroughly validated in production. Other than an experienced scientist she is also very good at interpersonal relationships integrating well with the rest of the team. Looking forward to working again together in the future.
Dr Elena Hensinger-Schien
Saatviga and I worked in the same research group and during this time, she was a professional, goal-oriented and efficiently-working colleague, focused and achieving her goals on time. At the same time, she is very nice and helpful, making it pleasant to work with her.
Andrea Pierleoni
I worked with Saatviga for two years at Healx on advancing deep learning methods applied to biomedical knowledge graphs to discover new drugs . Saatviga has a deep understanding of deep learning practices, she delved on a completely new research topic, got a good grasp of the literature around it and delivered multiple solutions thoroughly validated in production. Other than an experienced scientist she is also very good at interpersonal relationships integrating well with the rest of the team. Looking forward to working again together in the future.
Dr Elena Hensinger-Schien
Saatviga and I worked in the same research group and during this time, she was a professional, goal-oriented and efficiently-working colleague, focused and achieving her goals on time. At the same time, she is very nice and helpful, making it pleasant to work with her.
Andrea Pierleoni
I worked with Saatviga for two years at Healx on advancing deep learning methods applied to biomedical knowledge graphs to discover new drugs . Saatviga has a deep understanding of deep learning practices, she delved on a completely new research topic, got a good grasp of the literature around it and delivered multiple solutions thoroughly validated in production. Other than an experienced scientist she is also very good at interpersonal relationships integrating well with the rest of the team. Looking forward to working again together in the future.
Dr Elena Hensinger-Schien
Saatviga and I worked in the same research group and during this time, she was a professional, goal-oriented and efficiently-working colleague, focused and achieving her goals on time. At the same time, she is very nice and helpful, making it pleasant to work with her.
Andrea Pierleoni
I worked with Saatviga for two years at Healx on advancing deep learning methods applied to biomedical knowledge graphs to discover new drugs . Saatviga has a deep understanding of deep learning practices, she delved on a completely new research topic, got a good grasp of the literature around it and delivered multiple solutions thoroughly validated in production. Other than an experienced scientist she is also very good at interpersonal relationships integrating well with the rest of the team. Looking forward to working again together in the future.
Dr Elena Hensinger-Schien
Saatviga and I worked in the same research group and during this time, she was a professional, goal-oriented and efficiently-working colleague, focused and achieving her goals on time. At the same time, she is very nice and helpful, making it pleasant to work with her.
Credentials
-
Sun Certified Java Professional
Sun Microsystems
Experience
-
Healx
-
United Kingdom
-
Biotechnology Research
-
1 - 100 Employee
-
Senior Machine Learning Scientist
-
Jan 2022 - Present
- Propose and implement SOTA drug prediction methods based on learning network diffusion profiles in a biomedical graph - R&D lead for knowledge graph reasoning to support evidence generation for pharma rationale - Cross functional involvement with product and engineering teams to convert successful POC efforts to an usable product - Continuosly improve knowledge graph completion and reasoning models using SOTA GNN and transformer models - Mentor junior ML engineers or scientists
-
-
Machine Learning Scientist II
-
Jan 2021 - Dec 2021
- Build novel ML models for path-based knowledge graph reasoning using contextual embeddings and evaluate with internal data sets - Building benchmark datasets for path-based reasoning - Involve in drug matching projects to find treatments with developed ML models assisting Pharma rationale - Involve in external collaborations for research and publish findings
-
-
Machine Learning Scientist I
-
Feb 2019 - Dec 2020
- Evaluation and analysis of state-of-the-art knowledge graph completion (KBC) methods based on deep graph embeddings, re-inforcement learning and rule-based learning on a biomedical graph to improve accuracy of predictions - Building novel graph reasoning methods based on multi-hop path extraction and ranking using learnt graph embeddings and transformer architectures.
-
-
-
University of Bristol
-
United Kingdom
-
Higher Education
-
700 & Above Employee
-
Postdoctoral Research Associate in Machine Learning
-
Oct 2013 - Jan 2019
Experienced working on two EU funded Projects ThinkBig(ERC Advanced Grant) and CompLACS(EU FP7) - Probabilistic graphical models for fact checking using Probabilistic Soft Logic (PSL) in inferring implicit knowledge from text - Worked on a series of projects involving natural language processing and text mining tasks, semantic parsing, sentiment analysis, entity extraction, triplet extraction, entity disambiguation and building knowledge graphs to study macroscopic patterns in news… Show more Experienced working on two EU funded Projects ThinkBig(ERC Advanced Grant) and CompLACS(EU FP7) - Probabilistic graphical models for fact checking using Probabilistic Soft Logic (PSL) in inferring implicit knowledge from text - Worked on a series of projects involving natural language processing and text mining tasks, semantic parsing, sentiment analysis, entity extraction, triplet extraction, entity disambiguation and building knowledge graphs to study macroscopic patterns in news data. - Built and optimized statistical phrase based machine translation models to translate text in european languages used to monitor global changes of public opinion towards speci c topics/entities. - Implemented online classifiers on different relevance scoring functions: relevance to different news topics such as sports, politics etc and also appeal to readers of different news outlets such as BBC, NPR, Seattle Times etc. - Implemented a two-layer learning representation by combining modular adaptive modules into a single learning system where an intermediate representation of the data was learnt (hidden layer) by supervised online learning Show less Experienced working on two EU funded Projects ThinkBig(ERC Advanced Grant) and CompLACS(EU FP7) - Probabilistic graphical models for fact checking using Probabilistic Soft Logic (PSL) in inferring implicit knowledge from text - Worked on a series of projects involving natural language processing and text mining tasks, semantic parsing, sentiment analysis, entity extraction, triplet extraction, entity disambiguation and building knowledge graphs to study macroscopic patterns in news… Show more Experienced working on two EU funded Projects ThinkBig(ERC Advanced Grant) and CompLACS(EU FP7) - Probabilistic graphical models for fact checking using Probabilistic Soft Logic (PSL) in inferring implicit knowledge from text - Worked on a series of projects involving natural language processing and text mining tasks, semantic parsing, sentiment analysis, entity extraction, triplet extraction, entity disambiguation and building knowledge graphs to study macroscopic patterns in news data. - Built and optimized statistical phrase based machine translation models to translate text in european languages used to monitor global changes of public opinion towards speci c topics/entities. - Implemented online classifiers on different relevance scoring functions: relevance to different news topics such as sports, politics etc and also appeal to readers of different news outlets such as BBC, NPR, Seattle Times etc. - Implemented a two-layer learning representation by combining modular adaptive modules into a single learning system where an intermediate representation of the data was learnt (hidden layer) by supervised online learning Show less
-
-
-
University of Colombo School of Computing
-
Sri Lanka
-
Higher Education
-
100 - 200 Employee
-
Instructor
-
Oct 2009 - Sep 2010
- Conducted Lecture series and Tutorials in Software Engineering and Communication Technologies. - Involved in Undergraduate Paper Marking for Bachelor of ICT and Bachelor of IT degree programs. - Involved in the Post-implementation phase as a Developer in the eHealth project “Vidusuwa”. - Conducted Lecture series and Tutorials in Software Engineering and Communication Technologies. - Involved in Undergraduate Paper Marking for Bachelor of ICT and Bachelor of IT degree programs. - Involved in the Post-implementation phase as a Developer in the eHealth project “Vidusuwa”.
-
-
Education
-
2010 - 2014
University of Bristol
Doctor of Philosophy - PhD, Engineering Mathematics and Artificial Intelligence -
2005 - 2009
University of Colombo School of Computing
BSc, Information and Communication Technology
