Sr. Artificial Intelligence Engineer
Vebuin Co. Ltd.
Keywords
Artificial Intelligence
Amazon Web Services
Code Review
Distributed Systems
Python (Programming Language)
Node.Js
TypeScript
Cloudformation
Integration Tests
Restful Api
Skills
Python Backend, Typescript, AWS ECS, AWS Lambda, EC2, Redshift, AWS Redshift, RDS, S3, SNS, SQS, VPC, AWS, AWS services, AngularJs, Apache, API Gateway, API, Artificial Intelligence, AI, C#, cloud technologies, AWS CloudFormation, CloudFormation, Cloudwatch, code reviews, code review, Computer Engineering, Computer Vision, CI/CD, CouchDB, Data Streams, Deep Learning, distributed systems, distributed system, Django, Docker, Elasticsearch, Entity Framework, ETL, FastAPI, fault-tolerant, Flask, IAM, integration tests, integration testing, Jenkins, job scheduling, Kubernetes, load balancer, low latency, Machine Learning, Microservice, microservices architecture, mocha, MVC, Natural Language Processing, Nginx, Node.js, NodeJs, Numpy, object detection, OracleDB, programming languages, PyTest, Python, PyTorch, real-time data, Redis, REST API, REST APIs, SSIS, Scikit-Learn, Serverless Architecture, serverless, software development, software systems, Typescript, Websocket
Project history
10/2020
-
03/2022
- Owned the near real-time user behavior reporting and analytics product based on AWS Kinesis Data Streams,
Firehose, AWS Lambda, and Step functions . Ingests 4M+ user behaviors per day from ECommerce websites into
multiple AWS Redshift schemas with a serverless architecture, supporting a Metabase dashboard and multiple
KPI options for 10M+ users across multiple regions . Includes an AWS Lambda-based REST API that persists
streaming data into S3, supporting no-touch automatic backfills and minimal operational upkeep. The product
enables marketing managers and automated systems to make critical decisions for generating ECommerce
sales.
- Developed an End-to-End cost optimization solution for industries by utilizing Computer Vision, Rotated Mask
RCNN, Natural Language Processing, and microservices architecture deployed on AWS ECS, SNS, SQS , and
RDS . Implemented API Gateway and load balancer for server customization and reduced API latency time .
Delivered functionalities like invoice type detection, PDF split automation, and important entity recognition .
- Led the AI team and contributed towards the requirement gathering, user story designing, providing the inputs
in the infrastructure designing , solution development , Test cases designing , code reviews , and implementing
the CI/CD to deliver the quality solution.
Firehose, AWS Lambda, and Step functions . Ingests 4M+ user behaviors per day from ECommerce websites into
multiple AWS Redshift schemas with a serverless architecture, supporting a Metabase dashboard and multiple
KPI options for 10M+ users across multiple regions . Includes an AWS Lambda-based REST API that persists
streaming data into S3, supporting no-touch automatic backfills and minimal operational upkeep. The product
enables marketing managers and automated systems to make critical decisions for generating ECommerce
sales.
- Developed an End-to-End cost optimization solution for industries by utilizing Computer Vision, Rotated Mask
RCNN, Natural Language Processing, and microservices architecture deployed on AWS ECS, SNS, SQS , and
RDS . Implemented API Gateway and load balancer for server customization and reduced API latency time .
Delivered functionalities like invoice type detection, PDF split automation, and important entity recognition .
- Led the AI team and contributed towards the requirement gathering, user story designing, providing the inputs
in the infrastructure designing , solution development , Test cases designing , code reviews , and implementing
the CI/CD to deliver the quality solution.
04/2019
-
10/2020
Stack Machine Learning Engineer
Vebuin Co. Ltd.
- Owned the generalized receipt entity recognition product for detecting the meaningful insights from the
uploaded receipts for reimbursement purposes. Identifying the type and subtype of the receipt using natural
language processing , total amount, and issued date using RCNN object detection library with Google vision
(OCR) to automate the process of reimbursement module of a customized workflow in Smartflow Product. Able
to reduce the significant amount of time of clients and account department through the AI solution developed
using PyTorch, Django REST APIs, Python , and Computer Vision .
- Contributed to the driver safety module development for one of the largest automobile manufacturing
companies in Japan. Through machine learning created a pipeline for prediction of driver safety alerts using
the information of whether, present driving situations, driver and vehicle condition, and provided 89.93%
accuracy on real-time data .
uploaded receipts for reimbursement purposes. Identifying the type and subtype of the receipt using natural
language processing , total amount, and issued date using RCNN object detection library with Google vision
(OCR) to automate the process of reimbursement module of a customized workflow in Smartflow Product. Able
to reduce the significant amount of time of clients and account department through the AI solution developed
using PyTorch, Django REST APIs, Python , and Computer Vision .
- Contributed to the driver safety module development for one of the largest automobile manufacturing
companies in Japan. Through machine learning created a pipeline for prediction of driver safety alerts using
the information of whether, present driving situations, driver and vehicle condition, and provided 89.93%
accuracy on real-time data .
01/2017
-
04/2019
Software Engineer
Persistent Systems Ltd.
- Led development of insurance domain modules within the BFSI vertical, including a high-volume insurance
claiming system processing 4M+ daily requests . Managed 1B+ records via Microsoft SQL Server and developed
an analytic tool with .NET MVC 5, Entity Framework , and SSIS ETL functionality to support report management
and processing of 10M+ records according to insurance legal requirements.
- Contributed to TCC Designer module, by developing REST APIs in .NET MVC 5 with Entity framework and
OracleDB , and handled BA operations of designing the customized workflows to deliver the initial design of the
project outline to the development teams.
- Participated in Persistent Technothon 2017 , and developed the Traffic signal recognition module for driverless
cars and achieved 72.48% accuracy on the unseen data and scored 2nd Rank . Using OpenCV3, Python,
Scikit-Learn, Numpy, AWS, Machine learning, and Deep learning technologies delivered the driverless car
module which can predict the signal color from the live images captured by the camera and cover all the
edge cases of live scenarios faced at various road conditions.
claiming system processing 4M+ daily requests . Managed 1B+ records via Microsoft SQL Server and developed
an analytic tool with .NET MVC 5, Entity Framework , and SSIS ETL functionality to support report management
and processing of 10M+ records according to insurance legal requirements.
- Contributed to TCC Designer module, by developing REST APIs in .NET MVC 5 with Entity framework and
OracleDB , and handled BA operations of designing the customized workflows to deliver the initial design of the
project outline to the development teams.
- Participated in Persistent Technothon 2017 , and developed the Traffic signal recognition module for driverless
cars and achieved 72.48% accuracy on the unseen data and scored 2nd Rank . Using OpenCV3, Python,
Scikit-Learn, Numpy, AWS, Machine learning, and Deep learning technologies delivered the driverless car
module which can predict the signal color from the live images captured by the camera and cover all the
edge cases of live scenarios faced at various road conditions.
05/2016
-
12/2016
Trainee Engineer
Volansys Technologies Pvt Ltd.
- Contributed to App bundle feature development for a distributed system of one of the largest display
manufacturing providers in the USA. Using NodeJs, AngularJs, Elasticsearch, CouchDB , Jenkins and Nginx
developed a robust solution for distributed infrastructure and targets 4M+ client devices residing in multiple
regions.
- Developed feature testing pipeline using mocha and nock modules for end-to-end integration testing .
manufacturing providers in the USA. Using NodeJs, AngularJs, Elasticsearch, CouchDB , Jenkins and Nginx
developed a robust solution for distributed infrastructure and targets 4M+ client devices residing in multiple
regions.
- Developed feature testing pipeline using mocha and nock modules for end-to-end integration testing .
Local Availability
Only available for remote work
