Aadil Khan Email & Phone Number

Mountain View, California, Us

secure compute

Mclean, Virginia, Us

- Leveraged DynamoDB, S3, Sagemaker and SQS to build a scalable ML Platform that allowed internal teams to develop AI features by providing them with the infrastructure to run training and inference workflows.- Shipped a custom document extraction feature that allows customers to train models customized to their datasets. Training instances were run in Sagemaker and inferences in kubernetes to handle high volume.- Performed outage-less migration of an internal service from being deployed with gitlab runners to ArgoCD thereby saving about 3 hours worth of manual effort per week to deploy the service.- Identified the gap between the current performance metrics and the SLAs required to secure a 7 figure prospect by designing and running performance tests. Recognized bottlenecks and narrowed down the major areas to focus on to bridge this gap. Designed solutions for these problems and iteratively released improvements to get Appian's ML platform closer to the required SLOs.- Solved numerous customer incidents within the set SLAs and ensured that the customer's confidence in Appian's AI capabilities is not degraded.

Mclean, Virginia, Us

- Improved Appian's document extraction capabilities by switching from a native OCR model to AWS Textract to support use cases like skewed, handwritten and noisy text.

Jacksonville, Fl, Us

- Developed micro-services in Java and Python for a virtual assistant that processes mortgage documents by following agile and test driven development practices.- Enhanced an end to end system that redacted NPI information to support documents irrespective of their rotation and orientation.- Incorporated multi-threading to speed up services resulting in performance improvement of 35%

Jacksonville, Fl, Us

- Worked on an AWS based application that performs OCR, Classification and Extraction on mortgage documents. Translated data obtained from OCR and stored it in S3 which is consumed by the front-end.- Combined documents into packages based on configuration specified by customers.- Restructured data extraction entities of AWS Neptune and wrote queries for it using Gremlin to facilitate tracking of production failures to a specific entity.- Deprecated calls to older services, thereby reducing repeated production failures.

Master'S Degree, Computer Science

Bachelor Of Engineering, Computer Engineering