Summary
Overview
Work History
Education
Skills
Accomplishments
Affiliations
Certification
Timeline
Generic
Dasaprakash Krishnamurthy

Dasaprakash Krishnamurthy

Principal AI And Data Architect
London

Summary

Principal AI and Data Architect with 15+ years of experience across wide range of technologies in Big Data Analytics, Machine Learning Insights, Applied Computer Vision, NLP/NLU, Microservices Architecture, Cloud and DevOps. Worked in different roles as AI/ML Architect, Principal Data Scientist and Engineering Leadership. Have experience in end to end product and platform implementations right from architecture, design, implementation and delivery of large scale, highly available, distributed systems. Experience in architecting Big Data and AI/ML solutions that include Document Understanding AI, Deep Learning / Data driven pipelines, Call Centre Sentiment Analytics, MLOps. Exposure to product development lifecycle and working with top technology and consulting companies across geographically distributed teams with Agile Development practices. Certified in Deep Learning specialization from deeplearning.ai and reinforcement learning from HuggingFace.

Overview

15
15
years of professional experience
4
4
years of post-secondary education
1
1
Certification

Work History

Principal AI and Data Architect

UST
London
07.2019 - Current
  • Designed the Intelligent Document processing pipeline architecture using Generalized Task Orchestration principle. This generates and orchestrates a composable workflow dynamically using JSON configuration from creating a basic AI component to complex business process. Achieved reduction of capability onboarding time by removing engineering dependency
  • Designed highly scalable, distributed and performant production AI platform architecture to make customer and capability onboarding faster. This enabled research scientists to develop and productionize ML models on the go.
  • Implemented MLOps for the entire platform which involves Bring Your Own Model(BYOM), model training reusable pipelines, high-performance model serving, integrating with various data sources and bringing in multiple evaluation metrics
  • Implemented, Invented and filed for a revenue-generating patent for document understanding. This implementation extracts meaningful information from semi-structured and unstructured documents with text, tables and Barcode / QR Code by applying novel methods in AI like context awareness and information enrichment. This approach resulted in obtaining these insights by training only a few documents increasing throughput and accuracy and reducing cost for the platform
  • Implemented, Invented and filed a patent for straight-through processing using memory-enabled framework supported by reinforcement learning to identify highly accurate document templates to extract information and perform continual learning
  • Implemented, Invented and filed a patent for unified document classification pipeline which classifies documents by image instead of text. This unified classification pipeline can be applied across multiple document classification problems like document type, document relevance, document quality and document similarity as a no-code component
  • Designed and implemented the productivity tools like point and select and table rebound for human in the loop thereby improving time taken for review and enabling continual learning
  • Involved in forging partnerships, sales discussions, product strategy, requirements analysis, evaluating products, scaling platform with third-party offerings and publishing in cloud marketplace
  • Heading collaborative consultation with Stanford AI Labs (SAIL) to solve NLP and Computer Vision problems and participating in AI hackathons

Senior Data Architect

Cognizant Technology Solutions
08.2016 - 07.2019
  • Architecture of risk assessment platform for Small Commercial Insurance by sourcing data from disparate sources like ADLS/Databricks to Cosmos DB and gaining insights using AI/ML to provide instant quote
  • Digital real-time claims processing for baggage loss/ medical claims/ flight delay with Big Data Architecture that includes Kafka, Spark Streaming, Spark and Azure Cognitive Services
  • Actionable Objective Identification of Processes for RPA using sequence generation algorithms from recorded user actions in Spark using MLLib
  • Classification, Information Extraction and Knowledge Graph generation from multiple insurance processes like Commercial Insurance claims payment, Commercial Insurance Policy Registration and Renewal, Global Policy Clearance Compliance using Computer Vision, Natural Language Processing, Natural Language Understanding
  • Special Events Planner Websites Integration by creating a consumer digital channel by Integrating with Azure Bot Service, Azure APIs, Payment APIs and TPA APIs to provide instant quote
  • Call Centre Sentiment Analysis for near real-time analysis of call quality/guidelines and provide alerts/thresholds/suggestions based on sentiment, personality traits and emotional footprint of customer and agent using IBM Watson Cognitive Services
  • API Migration to Azure, Payment Gateway and Box Drop-in Integration, DevOps implementation

Senior Architect

Hexaware Technologies
12.2014 - 07.2016
  • Solution and Application architecture and design for Microsoft and Big Data applications
  • Architected ML and Data pipelines for clustering documents, assisted resolution of similar tickets, cheque classification
  • Led the creation of cluster models to optimize 5000+ Lotus Notes DBs(metadata) to under 40
  • Involved in reducing Technical Debt of suite of applications in a portfolio and implementing CI/CD with TFS
  • Architecture Definition, Assessment and Evaluation
  • Built competencies and mentoring
  • Defined technology strategy and roadmaps for the practice
  • Built technology assets – case studies, CoE strategy, Solution approaches, Technology Evaluation, Comparison Matrix, Identifying Trends, Frameworks and Tools
  • Lead and mentored Architects and Developers as part of projects and practice
  • Involved in pre-sales activities

Technology Architect

Infosys
12.2007 - 11.2014
  • Front-ended legal SaaS product design / implementation and deployment in Cloud including search engine of structured and unstructured content
  • Experience in designing System & Application Architecture, Pre-sales, Estimation, Development and Delivery of software products
  • Was part of the design team of a legal product company that covers Content Management Platform, Service Bus interface for distributed applications, Legal ERP solution, Search Engines over SaaS platform, Plug-ins for office components and browsers.

Education

Bachelor of Engineering - Computer Science & Engineering

Madras University
01.1999 - 04.2003

Skills

AI/ML: Deep Learning, Reinforcement Learning, Natural Language Processing

undefined

Accomplishments

Published Patents

• Neural Network Architecture for Extracting Information from Documents - https://patents.justia.com/patent/20220230013

Filed Patents

• Unified multimodal image-based document classification pipeline with context awareness

• Straight Through Processing through Memory Enabled Modelling and Reinforcement Learning

Affiliations

Stanford AI Lab

Certification

Deep Learning Specialization (deeplearning.ai) • Certification URL: https://www.coursera.org/account/accomplishments/specialization/R9GE9XXY5LKX

Timeline

Principal AI and Data Architect

UST
07.2019 - Current

Senior Data Architect

Cognizant Technology Solutions
08.2016 - 07.2019

Senior Architect

Hexaware Technologies
12.2014 - 07.2016

Technology Architect

Infosys
12.2007 - 11.2014

Bachelor of Engineering - Computer Science & Engineering

Madras University
01.1999 - 04.2003

Similar Profiles

CREATE PROFILE
Dasaprakash KrishnamurthyPrincipal AI And Data Architect