Dhaval Salwala

I am a Data Science Enthusiast exploring the power of Analytics and Machine Learning.

Biography

Currently, I am working in the position of Senior Research Engineer at IBM Research Ireland on a European Innovation Framework H2020 project on the Renewable Energy Sources (RES) sector.

Previously, I have worked with Dr Edward Curry in Multimedia Complex Event Processing in the Smart Cities and Sustainable IT group at the Insight Centre for Data Analytics at NUI Galway (NUIG). Here, I was researching in the field of real-time analytics and pattern detection for Multimedia Data Streams.

I am an experienced software professional with 7+ years of diversified industry experience. Led projects on developing ML solutions for multi-modal analytics using Computer Vision and Neuro Symbolic AI. I have worked on Computer Vision problems, multi-modal data analysis, Knowledge Graphs, Complex Event Processing and Deep Learning. Experienced in MLOps / AIOps processes including ML training / serving pipelines, CI/CD pipelines.

Passionate about building computer vision AI tools with research interests in Big Data and Reinforcement Learning (RL).

Interests
  • Machine Learning
  • Complex Event Processing
  • Multimedia Analytics
  • Reinforcement Learning
Education
  • MSc. Computer Science - Data Analytics, 2019

    National University of Ireland, Galway

  • BTech. Information Technology, 2012

    Gujarat Technological University, Ahmedabad

Experience

 
 
 
 
 
IBM Research Europe Dublin
Research Engineer - ML and Big Data
Jan 2022 – Present Dublin, Ireland
 
 
 
 
 
Data Science Institute, NUI Galway
Research Engineer - Multimodal Analytics
Oct 2019 – Present Galway, Ireland
  • Designed and implemented Computer Vision Pipeline for Gnosis MEP engine in Python, Redis Streams, Docker.
  • Assisted in scientific research for building Complex Event Patterns (CEP) in Data Streams. Worked on individual problem sets.
  • Worked alongside Research Fellows, Postdoctoral Researchers and PhD students on design challenges and refining CEP work problems.
  • Collected and studied large scale video datasets. Performed manipulating, processing, and extracting value from these datasets. Ran Image Processing models to analyse the accuracy and prepared ground data - TensorFlow/Keras.
  • Adopted efficient microservices architecture to execute ML pipeline on the streaming multi-modal data.
  • Designed benchmarking platform for a CEP Pipeline that includes Object Detection, Object Tracking and Annotation via Jaegar/Python.
  • Modelled data obtained through the CEP pipeline into a Knowledge Graph through spatiotemporal relationships.
  • Designed a matching engine for video querying that uses an openCypher query to explore Knowledge Graphs.
  • Paper on Query Driven Urban Accessibility using complex event processing accepted in ACM Multimedia 2021.
  • Assistant Chair and co-organiser for the 27th AIAI Irish Conference on Artificial Intelligence and Cognitive Science, Galway, Ireland, December 5-6, 2019.

Technologies: Python3, TensorFlow, Pytorch, Docker, RedisGraph, RedisStreams, DNN models.

 
 
 
 
 
  • Member of the Design Team that created technology stack for a risk interface application that monitors risk exposure in real-time for Eurex Exchange T7.
  • Led a team of associates to critical deliveries. Leveraged strong expertise in solving and identifying bugs while working with production support.
  • Worked closely with Business Analysts in mapping the business requirement with the Physical Data Model.
  • Developed AMQP interfaces to different cross-platform business entities and facilitated data exchange using Google Protocol Buffer (GPB).
  • Architect a state-of-the-art Cash Payment System (CPS) for Eurex Trading Platform T7. It offers a high degree of flexibility in terms of processing trades, payment locations, message formats and cross-currencies transactions. Created chained transaction handling using Spring Transaction Management.
  • Developed SWIFT message simulator in Spring Boot that performs load testing on thousands of concurrent payment instructions.

Technologies: Java8, EJB3, Python3, Spring Boot, Hibernate, Apache Qpid, Jboss7, PostgreSQL, Git, Maven, Protocol Buffers.

 
 
 
 
 
TCS - ICG Credit Approval System – Citigroup Corporation, NJ USA.
Software Developer
Mar 2013 – May 2015 Pune, India
  • Gained functional knowledge while working with legal, compliance and lending team on credit approval workflows and lending practices.
  • Optimized Citigroup’s Risk Management Business workflow using TIBCO iProcess and Spring Scheduler.
  • Enhanced UI response rate and operational efficiency of the platform by eliminating legacy codebase.
  • Working on Design Patterns, Unit Testing, Refactoring, Code structuring.

Technologies: Java1.6, Adobe Flex, Spring, Hibernate, IBM Websphere, Oracle 10g, Git, Maven, HP Load Runner.

Thesis

Msc. Data Analytics

.js-id-Multi-Agent-Systems
Master's Thesis

Master’s Thesis

EVALUATION OF MULTI-AGENT DEEP REINFORCEMENT LEARNING ALGORITHMS IN THE PURSUIT-EVASION ENVIRONMENT.

Research Projects

Data Science Institute

*
Vision-Farm

Vision-Farm

Query-driven Internet of Multimedia Things Platform for Smart Agriculture related Event Patterns

GNOSIS

GNOSIS

A Multimodal Event processing engine to Analyse Complex Event Patterns in Real Time

Academic Projects

NUI Galway

*
FAST VOTE AGGREGATION SCHEME

FAST VOTE AGGREGATION SCHEME

DAWID SKENE - FAST VOTE AGGREGATION SCHEME

GOVERNMENT ANALYTICS

GOVERNMENT ANALYTICS

MINING PAGES ON SOCIAL NETWORKING WEBSITES

Multi-Agent Deep Reinforcment Learning

Multi-Agent Deep Reinforcment Learning

Multi-agent implementation of the popular Reinforcement Learning algorithms.

ONE SHOT FACE RECOGNITION IN KERAS

ONE SHOT FACE RECOGNITION IN KERAS

Built a LeNet5 based deep convolutional model in Keras with a triplet loss function

Reinforcement Learning

Reinforcement Learning

Solutions to the most popular Reinforcement Learning algorithms.