I'm a developer with a passion for creating futuristic web experiences.
View My WorkOrchestrated the technical merge of developer documentation tooling with several new features, such as language-agnostic server-side SDKs for over 8 languages and an AI code copilot, increasing downloads in the updated portal by over 160%.
Developed a Jenkins pipeline to deploy API specifications from over 10 internal orgs through APImatic's Cloud API, deploying updates directly to CDN and webserver, enhancing deployment speed by 50% and reducing downtime by 30%.
Engineered a robust front-end portal utilizing React for dynamic rendering and GraphQL for efficient data fetching, tested using Cypress and Mocha with A/B testing by over 1,200 active users, resulting in a 25% uplift in daily user engagement
Java
JavaScript
Node.js
React
GraphQL
APImatic
Jenkins
CDN
Cypress
Mocha
A/B Testing
Docker
Kubernetes
Migrated 5 broadcast and communication microservices from Pivotal Cloud Foundry redis cache to Google Cloud Platform, improving scalability and leveraging real-time cloud-native services for enhanced performance and reliability.
Communicated with over 10 databases, utilizing RabbitMQ for seamless database amendment requests, for over 200,000 orders, performing conclusive validation on each, resulting in improved data integrity and increased speeds of up to 2x.
Modernized SOAP-based software architecture to RESTful API microservices, adding authentication security, Jacoco and Cucumber testing, and utilizing modern frameworks like Springboot, resulting in improved system performance by 40%.
Leveraged CI/CD Tekton pipeline for seamless deployment to preprod environments, ensuring efficient software delivery and integrating 42Crunch into development workflow for comprehensive security assessments and reliability.
Java
Spring Boot
Pivotal Cloud Foundry
Google Cloud Platform
Redis Cache
RabbitMQ
SOAP architecture
RESTful API
Jacoco
Cucumber
Terraform/Tekton
42Crunch
Built back-end software for real-time behavioral interview analysis, incorporating facial emotion tracking and sentiment analysis using the Facial Expression Recognizer library and DeepGram API.
Optimized user-facing video processing delays by 80% through techniques such as down-sampling and frame- skipping, while rendering content dynamically through client-side multithreading, leading to a 50% reduction in loading times.
Integrated facial expression recognition and speech-to-text conversion into the web app, enabling over a 1,000 monthly active users to receive instant feedback on expressions and analyze speech content for improved interview preparation.
Python
Facial Expression Recognizer
Deepgram API
SpaCy (NLP)
JavaScript
Client-side Multithreading
Speech-to-Text
Collaborated with a team to identify and alleviate software bottlenecks, improving application responsiveness by 5x.
Created over 200 mini-projects using a market-focused approach to streamline development.
Java
C#
Unity
Wwise
Microsoft Azure
JavaScript
HTML
CSS
Directed the entire product lifecycle from ideation to support, including material design and hardware integration.
Partnered with non-profits to provide a smart device for water filtration and usage monitoring on iOS and Android.
Collaborated with engineers using CAD for design components, facilitating testing for over 2000 users.
Python
Swift
Kotlin
CAD
Flow Rate Sensor
Bluetooth
Arduino Uno
AWS
JavaScript
HTML
CSS
Implemented the backend system to generate, store, and retrieve efficient routes for 65 athletes on the University of Michigan Track Team with five other developers.
Interacted with Google Maps API to consume routes and location data, transported data between frontend/backend via JSON, placed data in persistent storage (Firebase).
Developed API endpoints usable across iOS and Android platforms to enhance training plans by allowing athletes to analyze their routes and performance.
Designed full-stack social media web app for shortform video and launched beta with A/B testing on over 20 users
Engineered a RESTful API with robust security measures, including user authentication, data encryption, and protection against common web vulnerabilities, while using client-side dynamic pages
Formulated ML strategy aimed at enhancing urban green spaces across Prague utilizing Scikit-Learn, TensorFlow to analyze environmental data and identify optimal locations for green space development.
Integrated complex datasets on air quality, precipitation, and vegetation from active city sensors to create models that forecast the environmental impact of potential green projects.
Implemented a relational database with SQL syntax using C++, without STL data structures.
Supported operations like CREATE, INSERT, JOIN, REMOVE, PRINT, and GENERATE INDEX.
Explored hash tables through the generate index function.
Built a web-based office hours queue used at the University of Michigan.
Created a dynamic backend with linked lists and iterators to process API requests.
Enabled professors and TAs to assist students effectively with programming assignments.
Developed a weather app using data from OpenWeather API.
Focused on delivering information in a clean, modern interface.
Conducted extensive user interviews to refine design and features.
Created a classifier using the NLP Bag of Words model.
Determined thread or subject-matter of posts on Piazza.
Implemented a simplified version of the Multivariate Bernoulli Naive Bayes Classifier.
Feel free to reach out to me via email at krachit@umich.edu or follow me on social media.