Priyal Jain

Software Engineer

Hyderabad, Telangana, India4 yrs 11 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Stackforce AI infers this person is a Backend-heavy Fullstack Developer with expertise in Healthcare and SaaS.

Mern StackSaml 2.0Python (programming Language)MongodbDeep LearningConvolutional Neural Networks (cnn)

Data AnalysisData CleaningElasticsearchEncryptionNeo4jNode.jsPHPPandas (Software)UNET

4 roles

Apr 2025 – Present · 11 mos

Oct 2022 – Mar 2025 · 2 yrs 5 mos

Integrated AppDirect marketplace into the application and added recorded contact leads into Hubspot using Workato.
Created an Identity provider service to authenticate logged-in users on the marketplace using simpleSAMLphp.
I gained experienced in designing and integrating software systems using various technologies and frameworks for Monolith and Microservices such as python-pandas, ML models, husky, react, express, and rector.
I also worked on POCs for API gateway and authentication methods including Oauth2, JWT, SAML protocol, as well as data cleaning and analysis.

PHPNode.jsMERN StackSAML 2.0MongoDB

Jun 2021 – Sep 2022 · 1 yr 3 mos

Implemented field-level PII encryption and localization for multi-tenant level architecture using pre-save and post-save webhooks on microservices.
Developed a job scheduler algorithm on cost segregation on the mongo cluster efficiently based on time required for querying and data present for that tenant.
Created a microservice that fetches job descriptifon details using sentence-BERT open-source model and Neo4j database for mapping

Python (Programming Language)Pandas (Software)PHPNode.jsEncryptionElasticsearch+2

Jan 2021 – May 2021 · 4 mos

Jan 2020 – Dec 2020 · 11 mos · Hyderabad, Telangana, India · On-site

1. Mitosis count in WSI images to detect breast cancer:
Used a dataset consisting of MITOS-12, MITOS-14, and andrewjanowczyk images of varying quality and size.
Employed various preprocessing techniques and a segmentation model (Res2unet) to obtain uniform data and their respective ground truth.
Achieved higher accuracy of 82% using Detectron2 by Facebook.
Experimented with private dataset to synthesize the quality using cycleGAN to equalize accuracy with public dataset.
2. Detection of lesion in mammogram using Faster-RCNN:
Trained a model on a publicly available mini-MIAS dataset to detect the presence of a lesion in mammograms.
Achieved an accuracy of 68% using Faster-RCNN.
Research paper was published for this project.