Vivek Keshore

Green Street - Green Third Party

Green Street relies on a range of data pertaining to commercial properties and transactions for its commercial real estate intelligence and data analytics operations in both the US and EU regions. This data is procured from various third-party providers/vendors.

These vendors supply different types of data at varying frequencies, such as daily, weekly, monthly, bi-weekly etc. Earlier, the process of collecting, validating, transforming, and ingesting this data was done manually, requiring significant time and effort.

The GREEN project was aimed at addressing this issue and devising a solution with the end user in mind. The challenge was to eliminate manual intervention while ensuring the system remained user-friendly. The ultimate users of this system are analysts who utilize this data to deliver services to Green Street's customers.

Developed a cloud native solution using AWS, Airflow and Python to develop new ETL pipelines that would automate the process of data gathering, transforming and processing the third party data from all vendors. Also, developed a fully reusable and config based customised framework that would integrate any new or existing vendor data in Green Street system.

Skills: Python (Programming Language) · Apache Airflow · Amazon Web Services (AWS) · MySQL

Green Street - Data API One

I spearheaded a groundbreaking cloud-native project aimed at revolutionizing investment analyses in the commercial real estate sector. Leveraging cutting-edge technologies including AWS, FastAPI, SQLAlchemy, MySQL, Redis, and Pytest. Developed a robust platform from scratch to integrate Green Street's unparalleled commercial real estate data.

Developed a completely cloud-native solution that provided access to Green Street's best-in-class data, comprising over 4,000 data series accessible through 120+ endpoints across U.S. and Pan-European REIT and Data & Analytics products. This comprehensive dataset encompassed vital metrics such as NAVs, earnings, cap rates, IRRs, market grades, forward NOI projections, and more.

This application empowered the customers to incorporate Green Street's data directly into their underwriting models, asset management platforms, and data science applications. We streamlined the process by enabling automation of queries, ensuring that clients could access the most recent information programmatically or on-demand.I spearheaded a groundbreaking cloud-native project aimed at revolutionizing investment analyses in the commercial real estate sector.

Leveraging cutting-edge technologies including AWS, FastAPI, SQLAlchemy, MySQL, Redis, and Pytest. Developed a robust platform from scratch to integrate Green Street's unparalleled commercial real estate data. Developed a completely cloud-native solution that provided access to Green Street's best-in-class data, comprising over 4,000 data series accessible through 120+ endpoints across U.S. and Pan-European REIT and Data & Analytics products.

This comprehensive dataset encompassed vital metrics such as NAVs, earnings, cap rates, IRRs, market grades, forward NOI projections, and more. This application empowered the customers to incorporate Green Street's data directly into their underwriting models, asset management platforms, and data science applications. We streamlined the process by enabling automation of queries, ensuring that clients could access the most recent information programmatically or on-demand.

Skills: Python (Programming Language) · FastAPI · Amazon Web Services (AWS) · MySQL · pytest

Green Street - Phoenix

GreenStreet has 20+ code repositories that were having hundreds of sync/async tasks running periodical and non-periodical. The project aimed to modernize the task management infrastructure by transitioning from a customized Python-based Celery and Flower tool to Apache Airflow.

The tool had some disadvantages like role-based access, adding/editing/removing tasks, common dashboard for all repositories tasks etc. Also it was not scalable, not user friendly, and complicated to enhance. To eliminate all the disadvantages and scalability issues, the Green Street decided to use the Apache Airflow, and migrate all tasks to Airflow. Migration to Apache Airflow involves rewriting and re-implementing all the 245+ tasks in Airflow DAGs.

Apart from the Apache Airflow implementation of new DAGs in each repo, the team also worked in parallel on some APIs involving complex mathematical and GeoSpatial calculations.

The team did the RnD and developed a cloud-native solution using AWS MWAA, ECR, ECS and S3 along with automated CI/CD pipelines. The team came up with 4 different solutions along with sample PoC and demonstrated to the client along with the advantages and disadvantages of each approach/solution. Developed a standard scaffold to simplify the process of Airflow in each code repository.

GreenStreet has 20+ code repositories that were having hundreds of sync/async tasks running periodical and non-periodical. The project aimed to modernize the task management infrastructure by transitioning from a customized Python-based Celery and Flower tool to Apache Airflow. The tool had some disadvantages like role-based access, adding/editing/removing tasks, common dashboard for all repositories tasks etc.

Also it was not scalable, not user friendly, and complicated to enhance. To eliminate all the disadvantages and scalability issues, the Green Street decided to use the Apache Airflow, and migrate all tasks to Airflow. Migration to Apache Airflow involves rewriting and re-implementing all the 245+ tasks in Airflow DAGs. Apart from the Apache Airflow implementation of new DAGs in each repo, the team also worked in parallel on some APIs involving complex mathematical and GeoSpatial calculations. The team did the RnD and developed a cloud-native solution using AWS MWAA, ECR, ECS and S3 along with automated CI/CD pipelines.

The team came up with 4 different solutions along with sample PoC and demonstrated to the client along with the advantages and disadvantages of each approach/solution. Developed a standard scaffold to simplify the process of Airflow in each code repository.

Skills: Amazon Web Services (AWS) · Python (Programming Language) · Apache Airflow · FastAPI · MySQL

The RISCS Ratings System^TM

The RISC Rating System^TM is the first and only platform that uses non-public manufacturer-reported data to provide robust supply chain analytics at the NDC level.

The RISC Ratings Portal enables pharmaceutical manufacturers and healthcare professionals from group purchasing organizations and health systems to submit, edit, or view supply chain data for critical pharmaceuticals.

The system is developed to help GPOs or Hospital or Drug Manufacturers to know more about the availability and rating of the drug. The rating is awarded based on some complex computation which takes into consideration several factors like number of raw material supplier, geographic distribution, stock, shelf life and so much more.

Field Pros Direct

Field Pros Direct is an automated assignment platform that matches experienced adjusters and service providers with claims based on customized preferences. A tech-enabled claims solution to assist every aspect of the claims process - from start to finish.

Field adjusters, desk adjusters and other claims service providers are matched with assignments based on expertise. The smart-matching method helps ensure accurate indemnity, while eliminating litigation expenses.

Field Pros Direct brings all service providers under one umbrella, and provides an on demand claim adjuster, and thus facilitating everything end to end from Servicer Provider, Adjuster and the end customer perspective. The system is seamless and automated. Automated claim assignment, appointment scheduling, automated communication through emails and sms, automated report generation, claim settlement, invoice generation and much more.

HBK Capitals

Development and design of a data processing engine for a US based Investment Banking client. The DImE engine is built in a robust way which would connect to number of third party data providers (free and paid) to collect the data which would be used by client to make trades. Various sources provide data in different file format and data format. DimE engine is built to process, normalize and analyze the data.

Zonar Systems

Development of a web based application from scratch for Zonar Systems in Python, Pyramid and other related technologies. The role also involved taking part in design discussions and data models. The application was to control and monitor the live traffic of ground based fleet of vehicles. It was similar in a tone like air traffic control system, but for ground based fleet.

Flask API Starter Scaffold

Gitlab code - Flask API Starter Scaffold

A starter kit which would have everything configured and built in it. So, that any new project can be started within minutes, or any PoC can be developed as quickly as possible. The scaffold will make sure that each project would follow same standards and structure. The scaffold comes ready made with features like User Management, Authentication and authorization, Security & Encryption, Caching, Logging, Swagger, Celery, Celerybeat, Celery Tasks, Alembic, Docker, Layered architecture, custom database seed & much more.

SG Aspire

Talent Acquisition team members have to go through a lot of manual efforts in order to search and shortlist the candidate’s resumes which they receive from multiple sources. This is challenging, time consuming and laborious. This solution automates the process of collecting the resumes from multiple sources, and parse it, so that the shortlisting process can become more efficient.

The developed automated system reads the emails and downloads the resumes received from candidates. Downloaded resumes are automatically stored central & secured OneDrive repository.

The system also reads and parses the resumes, and extracts all the relevant information like phone, email, skills, experience etc.

The application automatically acknowledges each and every candidate, and also tracks the progress of interview process of candidates.