Joshal Daftari Email & Phone Number

Mountain View, California, Us

Prague, Cz

Prague, Cz

• Lead a team of local and remote engineers to accomplish several deliverables with aggressive deadlines⠀⠀⠀ * Reduce operational cost by 70% by migrating geocoding for all our apps from Bing Maps to Mapbox.⠀⠀⠀ * Pilot packet sniffing using Scapy to calculate NetFlow from multiple home IoT devices and aggregate it to detect vulnerability.⠀⠀⠀ * Incorporate sign-up and auth mechanism for Verizon Smart Family app by integrating with carrier platforms.⠀⠀⠀ * Architect an in-house push notification solution and migrated our mobile clients from PubNub for better reliability and debug capabilities.⠀⠀⠀ * Defined and improved data collection pipelines by using AWS Kinesis and Redshift.• Build fully automated CI/CD pipelines for containerized applications using Jenkins, Spinnaker, Kubernetes, Terraform, and AWS.• Evaluate feasibility of Datadog in a company Hackathon, which has now become an integral part of our monitoring stack.• Collaborate with product, design, and project teams to define high-level milestones and transform them into structured engineering tasks.• Plan low-downtime cross-team feature releases and work with the Operations team to deploy and debug on production servers.• Revamp our onboarding tutorials to use our latest tech stack, enabling new hires to contribute within the first week.• Mentor engineers by guiding them through code reviews to achieve the best development, testing, logging, and monitoring practices.

Emeryville, Ca, Us

• Transitioned our entire legacy stack to Ansible to reduce deployment times by 50% for all our parental controls products.• Added backend support for processing Apple In-App purchase for our Family Locator iOS apps.• Worked on a team that led API-first development effort and redesigned our backend monolith into microservice and event-driven architecture.• Implemented several horizontally scalable microservices in Python Flask and Java Spring Boot deployed in Docker containers.• Ensured product success and uptime by debugging customer issues and providing on-call production engineering support.

Emeryville, Ca, Us

San Diego, Ca, Us

• Automated periodic fetching, reporting, and parsing of logs generated at remote access points to provide a statistical insight on the overall system behavior.• Developed a man-in-the-middle CLI tool to listen, filter, or drop messages between system components.• Rectified and enhanced our messaging libraries to support end-to-end payload generation and verification.• Accelerated the backend regression test cycles by 80% by creating a simulator for third party applications.• Designed a REST client interface and XML data parser to facilitate debugging of our deliverable API.• Built tools to generate real-time statistics and visualizations to monitor our pilot system of 150 devices.• Implemented a client for Qpid AMQP broker to send user data derived through a load test model.

La Jolla, Ca, Us

mHealth Project • Part of a multidisciplinary group working on research, modeling, and analysis of biometric sensor data.• Developed a GUI for running data mining and signal processing algorithms and querying for patterns of well-being from heart-rate variability time-series. Designed an algorithm for finding synchronicity and causal relations among individuals in group activities like Yoga. • Formulated upper bound on the samples required to implement the method in real time with <5% error.• Analyzed the accelerometer data to search specific events for classifying human gait and checking for gait symmetry.• Made the tools portable to run on a backend engine for user physiological data analytics and visualization.• Continued to volunteer the effort even after graduating. The work resulted in 1 journal, and 2 conference publications.Technologies used: Python, Matlab

La Jolla, Ca, Us

TA to Prof. Robert Lugannani for the upper division course "Probability and Random Processes for Engineers"

San Diego, California, Us

Cognitive Networking• Implemented a machine learning technique to forecast the best configuration for a Wireless LAN system.• Worked with a team to develop and test the probabilistic graphical model, Bayesian Networks, on an embedded device for real-time traffic prediction.• Improved error performance, precision and throughput by 35%, 40% and 20% respectively.• Extended goals to publish a paper for an international conference as well as a demo event and presented the project at university research exposition.Technologies used: Matlab, C, MySQL, Shell scripting, R

San Diego, California, Us

• Devised a decentralized strategy for network nodes to choose a Wi-Fi channel adaptively using a Markov Random Field and Maximum A Posteriori framework.• Programmed the model for channel classification in Python using historical data to attain 95% accuracy.

Master Of Science - Ms, Electrical & Computer Engineering

Bachelor Of Technology - Btech, Electronics And Communications Engineering