portfolio@ryandielhenn:~$
drwxr-xr-x 2.1K Sep 14 about/
drwxr-xr-x 4.2K Sep 14 experience/
drwxr-xr-x 3.8K Sep 14 projects/
drwxr-xr-x 1.9K Sep 14 education/
-rw-r--r-- 245K Sep 14 resume.pdf
-rw-r--r-- 64B Sep 14 github.link
-rw-r--r-- 64B Sep 14 linkedin.link
-rw-r--r-- 64B Sep 14 email.link
portfolio@ryandielhenn:~$ _
portfolio@ryandielhenn:~/about$
About
M.S. Computer Science candidate specializing in scalable distributed systems and applied machine learning. My background includes contributions to Apache Kafka's transition to a self-managed metadata architecture (ZooKeeper removal), including work on Confluent Cloud.
Skills & Technologies
JavaGoCPythonApache KafkaDockerLinux
portfolio@ryandielhenn:~/experience$
Experience
My professional journey in software engineering and distributed systems.
Confluent
- Contributed to Apache Kafka's migration from ZooKeeper to KRaft, improving observability and reliability of the distributed consensus layer in Java/Scala.
- Built metrics pipelines in Java/Scala to monitor cluster health, quorum state, and inter-broker communication patterns for KRaft; updated Confluent Cloud tooling in Go to integrate new metrics.
- Contributed to Confluent Cloud's Cluster Linking integration with KRaft architecture for cross-cluster replication.
Confluent
- Implemented dynamic client reconfiguration in Java/Scala for Apache Kafka, enabling runtime updates to producer/consumer settings (including connection, security, retry, and ack configurations) without service restarts.
- Enhanced Confluent Cloud's rebalance tooling with asynchronous replica movement support.
- Continued contributing to Apache Kafka during Fall 2020 while completing undergraduate degree.
Big Data Lab - University of San Francisco
- Developed a C implementation of Geopresence, a bitmap-based geospatial indexing system, using RoaringBitmap compression and HyperLogLog++ for efficient location queries on IoT devices.
- Implemented C-based adaptive grid index achieving 17x speedup over Java; outperformed R-tree indexing by 400x at scale (1M+ points) while R-trees showed advantages on sparse datasets (<7K points).
University of San Francisco
- Designed projects and coursework, reviewed student code weekly, and provided feedback with grades.
- Held weekly office hours to guide students through project design and debugging.
University of San Francisco
- Supported the CS department by updating, patching, and maintaining Linux lab machines, improving system stability for students and faculty.
portfolio@ryandielhenn:~/projects$
Projects
A collection of my work in distributed systems, data science, and software engineering.
- An interactive EDA dashboard with FastAPI and DuckDB for browser-based analysis of million-row datasets with sub-second query performance.
- Implemented bias detection and fairness analysis to identify discriminatory patterns across demographic groups and data drift detection to flag distribution shifts.
- Leveraged DuckDB to enable fast and memory efficient analytical queries on datasets with millions of rows.
PythonFastAPIDuckDBData AnalysisBias DetectionFairness AnalysisState ManagementCachingData Visualization
- A distributed caching system in Go with consistent hashing for key distribution, achieving 49K ops/sec throughput (10K operations, 32 concurrent clients, 128-byte values).
- Includes an observability stack with Prometheus/Grafana to monitor request latency, throughput, and error rates, plus benchmarking tools to measure routing efficiency and performance under load.
GoDistributed SystemsConsistent HashingGossip ProtocolDockerPrometheusGrafana
- A bitmap-based geospatial indexing system designed for IoT and low-power devices using RoaringBitmap compression and HyperLogLog++ for efficient location queries.
- Developed during undergraduate research at University of San Francisco (January 2020 - December 2020).
- Implemented a C-based adaptive grid index that achieved 17x speedup over the Java implementation and outperformed traditional R-tree indexing by 400x at scale (1M+ points), while R-trees showed advantages on sparse datasets (<7K points).
- Features probabilistic point density estimation and adaptive grid resolution optimized for edge computing environments.
CC++JavaGeospatial ComputingIoTRoaring BitmapsHyperLogLog++Edge ComputingSpatial Indexing
- A distributed filesystem built with Java, Netty, and Protocol Buffers.
- Features a controller-based architecture that routes client requests to storage nodes using Bloom filters to minimize unnecessary lookups.
- Implements data chunking across storage nodes with replication for fault tolerance and availability.
JavaNettyProtocol BuffersBloom FiltersDistributed Systems
- A multithreaded search engine that tracks user queries, enables web crawling, and supports search over an inverted index built from the crawled pages.
JavaMultithreadingInverted IndexWeb CrawlingJetty
- A command line shell built in C with support for scripts, heredocs, pipes, redirection, background jobs, command history with bang reruns, and built-in commands like cd and setenv.
- Features arrow key navigation through history and input editing.
CSystems ProgrammingShell ImplementationProcess ManagementSignal HandlingPipesRedirection
portfolio@ryandielhenn:~/education$
Education
My academic background and achievements.
California State University, Los Angeles
- Specializing in scalable distributed systems and applied machine learning.
University of San Francisco
- Graduated Magna Cum Laude with a 3.75 GPA
- Minor in Mathematics