Eashaan Katiyar

Hi! My name is Eashaan

IMC Trading Juniper Networks Intel neurIOT UC Berkeley

Spearheaded the development of new features for PyramidGen, a high-performance low-latency project in C++ enhancing IMC's competitive edge in valuation-triggered trading.
Proactively automated tasks and optimized code generation in the PyramidGen codebase, boosting developer efficiency and code maintainability.
Overhauled the build process for the entire execution codebase, improving failure reporting, resolving flaky tests, and introducing network-level synchronization of multi-threaded tests. This initiative led to a 95% success rate in builds (up from sub-50%) and a reduction in build times, significantly accelerating CD and release turnaround while minimizing developer friction.
Served as team lead for the PyramidGen project, overseeing performance enhancement, bug fixing, and the onboarding of junior developers. Additionally, streamlined regression systems to enhance reproducibility and visibility.

Developed the back-end of the WAN Assurance project, streamlining corporate network operations and enhancing network uptime. This service is actively utilized by Fortune 500 companies
Integrated Juniper's firewall devices by implementing IPSEC features for the WAN-Assurance micro-services in Go and developed robust data pipeline using Apache Kafka & Storm, ensuring secure network communication.
Overhauled the micro-service codebase and associated protobuf structures to ensure compatibility with high availability firewall systems, enhancing the overall reliability and performance of the system.

Developed a Perl application that significantly reduced regression runtime and testing turnaround for Intel's 7 and 10nm server designs. Leveraged thread-level parallelism and distributed computing to execute granularized test suites
Revamped tools for communicating chip design readiness to customers, enhancing the efficiency and accuracy of communication processes.

Designed a machine learning application to forecast sales of prospective sunglass designs for a prominent sunglasses manufacturer in India
Applied OpenCV and scikit-learn to analyze sunglass fashion features from images, implementing advanced algorithms for feature extraction and image classification.

Work on an RL model to navigate rovers on extraterrestrial soil for exploration and resource collection
Simulated a rover as an agent within a Markov Decision Process with improvements through Proximal Policy Optimization using OpenAI Gym and simulated through ROS Gazebo

Mentor students in Electrical Engineering concepts and practice problems for the Designing Information Devices and Systems course
As a senior mentor, I helped new members of the club improve their teaching ability and additionally conduct content creation and review
Led a mock midterm exam and review session before a midterm, lecturing to a class of more than 100 students

Developed an Augmented Reality app and associated control and obstacle avoidance algorithms to communicate with a bare-metal robot over BLE and enable it to autonomously execute a user-defined path
Competed in a class algorithm design competition, for which I implemented multiplicative weights and steiner trees, and created a testbed to conduct hyperparameter training on the cloud. I placed in the top 10% of the class
3-man team created voice-controlled model car using PCA, SVD, and k-means, along with RC filters and closed-loop feedback control

Service that analyzes user listening behavior to add and remove songs, maintaining library & playlist quality

Python SQL SaaS Stats

Partnered with FakeNetAI to develop, train, and test Dilated Res-Net models to detect facial image manipulations using software like Photoshop.

Machine Learning PyTorch Data Science

Predicts and generates optimal daily fantasy lineups. Using it, I ranked in the top 5% of players in Yahoo DFS

JS LP Algo

Command-Line application that summarizes webcast lectures by removing unimportant segments

Python NLP AI