● Pioneered the architectural design and development of a customer-centric platform, utilizing a choreography of microservices communicating through an Apache Kafka message queue.

● Engineered a distributed inference service for large-scale virtual staining of unlabeled tissue samples via AI using Python and Torchserve.

● Coordinated the integration of multiple fluorescence and brightfield microscopes, including Zeiss (.czi), Akoya (.qptiff), 3DHistech (.mxrs),

Olympus (.orf) within both backend and frontend services, through the use of a unified interface.

● Embedded FDA-compliant audit trail functionalities to ensure data integrity.

● Engineered a Python service to perform Whole Slide Image registration as a service, using several image alignment algorithms such as Deep

Learning based feature matching, SIFT, and other key algorithms.

● Engineered a Python service to perform Whole Slide image dicomization (ISO12052) and storage to a DICOM server (Orthanc).

● Principal contributor to a dedicated service that monitors scan completions from microscopes and executes multi part uploads using a PyQT

client interface.

● Developed Software Design (C4) and Software Requirement Specifications, led Design Reviews, integrated design inputs and outputs within the company’s

QMS system.

● Piloted the transition of the main customer-facing API from Role-Based Access Control to Attribute-Based, using a Nest.js backend and Casl.js

● Ensured software compliance with ISO 62304, ISO 13485, ISO14971 and other key medical standards.

● Led and mentored a team of 5+ professionals, emphasizing workflow efficiency and strong quality assurance.

● Developed an open-source-integrated API for enhanced data matrix reading, achieving an 80% production read improvement and nearly 2x cost savings.

● Designed and implemented a data review tool using Rust, enabling precise identification of optimal fields of view within Whole Slide Images (WSIs)

● Played a pivotal role in shaping the company’s tech stack, working extensively with Typescript, Python, Rust, and C++.

● Managed customer engagements and spearheaded deployments across the U.S which performed thousands of scans without interruption.

● Coordinated the deployment of a scalable Kubernetes based cloud infrastructure to serve customers across the globe.

● Created a Python framework for neural network training with Pytorch Lightning, adopted company-wide. Incorporated advanced hyperparameter tuning

using Ray and Optuna.

● Achieved a 10x speed up on proprietary research models utilizing a TensorRT based, customized build of Torchserve.

● Led the design of a framework for efficient fluorescence/brightfield whole slide image ingestion and compilation into datasets for supervised learning.

● Developed, hosted, and maintained a React website highlighting the startup's innovative achievements.

● Conceived and executed the creation of a sophisticated framework to ensure the training and deployment of Keras models, utilizing Tensorflow Extended.

● Spearheaded the design and realization of an all-encompassing full-stack infrastructure to facilitate streamlined annotation of Whole Slide Images. This

system seamlessly integrates an Orthanc/Flask backend with a React-driven front-end visualization.

● Coordinated the deployment and maintenance of such systems on Amazon Web Services.

● Developed an AI model with PyTorch using Super Resolution Generative Adversarial Networks, analyzing the relationship between perceptual accuracy and metrics such as SSIM/PSNR.

● Devised a full stack solution (using React and Flask) enabling remote command of a Raspberry Pi, repurposed as a nimble, digital microscope.