Education

• Ph.D. in Computer Science, University of Houston, 2018
  • Dissertation: “Domain Adaptation using Deep Adversarial Models”
  • Committee: H. Ngueyen, W. Shi, A. Mukherjee (co-Chair), R. Vilalta (Chair)
• B.Sc. in Computer Science, University of Houston, 2012

Employment History

• 2019-present Senior Scientist, NLP Research, Aon IPS
• 2018-2019 Lead Data Scientist, SnapRapid, LLC.
• 2017-2019 CTO and Co-Founder, AI Savvy, Inc.
• 2017-2018 Data Scientist, AE Studio
• 2015-2017 GAANN Fellow/Research Assistant, Pattern Analysis Laboratory, University of Houston
• 2015-2017 Teaching Assistant, Machine Learning, AI, Advanced AI, University of Houston
• 2012-2015 Machine Learning Research Scientist, I2C Laboratory
• 2006-2012 Full-Stack Software Engineer, DomainParking LLC.

Internships

• HRL Laboratories (2 summers) - key research areas: Cloud Computing, Data Engineering, AI, Security
• Air Force Research Laboratory - key research areas: NLP, NLG, Social Networks

Technical and Management Experience

Aon

• Leadership, researcher, and individual contributor role in an NLP Research Lab
• DMajor projects elivered to production: 1) Initiated and contributed to the development of a sentence-transformers library 2) Contributed to the training of domain-specific BERT-like embeddings for document clustering
• Individual contributions or work with a small engineering team 1) Leading contributor to an in-house document classifier library 2) NER-based classifier and a toolkit to train on new labels and entities given an arbitrary grammar
• Research 1) Introduced, evangelized, and lead the training of deep language models and embeddings 2) Influenced numerous projects, providing a business contribution across the board 3) Trained in custom LSTMs and Transformers on domain-specific Big Data (# documents > BERT) 4) Designed and delivered several loss types to benefit from structured and image data available 5) Lead research in self-constructing knowledge graphs with deep multi-modal embeddings
• Provided business leadership with ideas and advice based on the latest developments
• Proposed solutions and present results to audiences varying from CEO and CTOs to large audiences
• Transformers (BERT, GPT), PyTorch, Tensorflow, AWS, GCP, Python, Pandas, Spacy, NLTK, Flair, FastAI

SnapRapid

Researched, designed, developed and delivered solutions that use NLP/Machine Learning to estimate influence of social media accounts featuring multi-media content. Used in a recommendation engine.

AI-Savvy

Responsibilities: Data Science, Solution Architecture, Business Development, delivery of product. Managed a technical team of 10, 2 PhDs and 3 MS Data Scientists. External interaction with CEOs and CTOs.

AE Studio

Researched and developed solutions for clients who needed Visual QA and Visual Dialogue. Used Image Recognition, NLP and Deep Learning to help achieve their goals.

Pattern Analysis Laboratory

Research in Machine Learning and NLP. Projects for oil & gas and NASA.

University of Houston

Responsible for applied & experimental component of several graduate courses. Assisted in instruction of theoretical machine learning concepts. Mentored several students.

I2C Laboratory

Machine Learning Research Scientist on an $5 million project to prevent DDoS attacks on 911 emergency channels. Lead a team of 7 employees on a $1 million project to provide machine learning based user verification. Reported to the Cyber Security Division of HSARPA.

DomainParking

Responsible for connecting front-end with the models delivered by the data analytics team.

Technical Skills

• Expert: Python, PyTorch, Keras, Tensorflow, fastai, sklearn, Pandas, NLTK, spaCy, Numpy, Unix, C/C++
• Experienced: MongoDB, PostgreSQL, MySQL, Neo4j, Plotly/Matplotlib/Seaborn, Scipy, AWS, Docker, Java
• Comfortable: R, Matlab, Caffe, Tableau, OpenCV, Spark, MLlib, 10+ programming languages

Projects

• openSMILE - Co-author of openSMILE for Android, popular software used in Acoustic Pattern Analysis. More than 10 downloads/day for 4 years.
• NG-911 DDOS Prevention: comprehensive software package that utilizes AI to prevent DDOS attacks originating from text, VoIP & other means. Solution currently being adopted by 911 Call Centers.
• Github: dainis-boumber

Publications

• Dainis Boumber, Yifan Zhang, Arjun Mukherjee, “Experiments with Convolutional Neural Networks for Multi-Label Authorship Attribution”, LREC 2018, 11th Conference on Language Resources and Evaluation
• Vilalta R., Dhar Gupta K., Boumber D., Meskhi M. M., “A General Approach to Domain Adaptation with Applications in Astronomy”, Publications of the Astronomical Society of the Pacific (PASP), 2018, IOP Science Press.
• Pablo Guillen (University of Houston), German Larrazabal (Repsol USA), Gladys González (Repsol USA) Dainis Boumber (University of Houston), Ricardo Vilalta (University of Houston), “Supervised learning to detect salt body”, 2015 SEG’s International Exposition and 85th Annual Meeting in New Orleans, Louisiana
• W. Shi, Y. Wen, Z. Liu, X. Zhao, D. Boumber, R. Vilalta and L. Xu, “Scalable and Fault Resilient Physical Neural Networks on a Single Chip”, CASES 2014
• Tao Feng, Zhimin Gao, Dainis Boumber, Tzu-Hua Liu, Nicholas DeSalvo, Xi Zhao and Weidong Shi, “USR: Enabling Identity Awareness and Usable App Access Control During Hand-free Mobile Interactions.” Mobicase 2014
• Lee, Z. Liu, X. Tian, D. H. Woo, W. Shi, D. Boumber, Y. Yan, and K. Kwon, “Acceleration of bulk memory operations in a heterogeneous multicore architecture”, PACT 2012
• T. Feng, Z. Liu, B. Carbunar, D. Boumber and W. Shi, “Continuous Remote Mobile Identity Management Using Biometric Integrated Touch-Display”, MICRO Workshops 2012: 55-62.

Awards

• Academic Excellence Award
• Dean’s List