kidus.sendeke
Seattle, WA

Software engineer · Seattle, WA

Kidus Sendeke

Distributed systems · AI-enabled apps · Founder, 50 Limes

I build reliable backend systems and the AI products that run on them. Today I build full-stack, AI-enabled applications for Microsoft Philanthropies with Blueprint Technologies. Before that I spent three and a half years on Microsoft's M365 Substrate team, working on ingestion APIs, global rate limiting and big-data pipelines, after starting my career at Visa.

See my work → Resume (PDF) ↗ kidussendeke@gmail.com
// 01

Now

updated Oct 2026
Work

Building full-stack, AI-enabled applications for Microsoft Philanthropies in C#/.NET, TypeScript, Azure OpenAI and the Bot Builder SDK.

School

M.S. in Computer Science, Data Science specialization, at Seattle University. Graduating Spring 2027.

Building

Natural World in Numbers, a measurement engine for phones and drones, and EthioData Hub, market intelligence for Ethiopia's opening economy.

Shipped

Anchor at HuskyHack 2026 (Sound Credit Union × UW Tacoma Milgard Center): protection for credit-union members against family-emergency scams and account takeover.

// 02

Experience

since 2018
Present
Current

Blueprint Technologies → Microsoft Philanthropies

Greater Seattle, WA

Software Engineer (contract)

  • Build full-stack, AI-enabled applications for Microsoft Philanthropies, from backend services to the user-facing experience.
  • Work across C#/.NET services, TypeScript front ends, Azure OpenAI and conversational interfaces built on the Microsoft Bot Builder SDK.
  • C#
  • .NET
  • TypeScript
  • Azure OpenAI
  • Bot Builder SDK
Nov 2020 — Apr 2024
3 yrs 6 mos

Microsoft

Redmond, WA

Software Engineer · M365 Substrate

  • Built a C# API for SharePoint data ingestion that connects three client systems and made user file processing 40% faster.
  • Raised the availability of a global rate-limiting service by 20% with cloud metrics and monitoring.
  • Ran a QPS analysis across thousands of data shards that found 15% under-provisioned, then added dynamic feeding so hot shards scale, improving throughput and resource allocation by 25%.
  • Developed a big-data pipeline in SCOPE that raised data-validation accuracy by 30%.
  • Added advanced SDK control features to .NET Core apps, improving the global user experience by 50%. This work led to my promotion.
  • Carried on-call for 5 critical distributed backend services (24/7 for 2 years, then 12/7 for 1.5 years) at 99.9% uptime.
  • Mentored two interns through complex projects. Both received return offers.
  • C#
  • .NET Core
  • SCOPE
  • Azure Data Explorer
  • KQL
  • Cosmos
  • Distributed systems
Sep 2019 — Oct 2020
1 yr 2 mos

Visa Inc.

Austin, TX

Software Engineer

  • Built Python and Selenium automation suites that cut testing time by 50% and manual effort by 60%.
  • Created regression and volume-testing procedures that improved product performance by 15%.
  • Engineered a Go utility for EBCDIC-to-ASCII conversion of mainframe files that halved processing time.
  • Designed Python classes for dynamic transaction-file generation, reaching 100% testing accuracy.
  • Python
  • Selenium
  • Go
  • Payments
Jun — Sep 2018
Internship

Visa Inc.

Austin, TX

Software Engineer Intern

  • Developed a Jira-like project-tracking web app in Angular that made tracking 30% more efficient for managers.
  • Built interactive dashboards with D3.js and Chart.js to visualize project and application progress.
  • Designed MongoDB backend schemas and created mock services for development.
  • Angular
  • D3.js
  • Chart.js
  • MongoDB
// 03

Projects

github.com/Kidusyohanes ↗
// p-01Deployed

LimeGate

Multi-provider LLM inference gateway

Routes LLM traffic to AWS Bedrock with Azure fallback behind a reliability layer: Lua-based token-bucket rate limiting, circuit breakers, PII redaction and cost attribution. Ships with SLOs and burn-rate alerts, runbooks, Argo canary deploys, a typed SDK, k6 load tests and Grafana dashboards.

  • AWS EKS
  • ElastiCache
  • Bedrock
  • Azure
  • Argo
  • k6
  • Grafana
// p-02Flagship · 50 Limes

EthioData Hub

Market intelligence for Ethiopia

Market intelligence for Ethiopia's fast-opening economy, built around the birr float, the IMF credit facility, the new Ethiopian Securities Exchange and the opening of the banking sector. Next up: a retrieval-augmented (RAG) research assistant on top of the data.

  • Data platform
  • RAG
  • Emerging markets
// p-03In progress

QuickLime

Low-latency trading system in C++

A high-frequency trading system built from the core up. Phase 0 is complete: benchmarked single-producer, single-consumer (SPSC) queue and memory-pool components, with a GitHub Actions pipeline that gates merges on performance regressions.

  • C++
  • Benchmarking
  • GitHub Actions
  • CI/CD
// p-04In progress

LimeCart

Event-driven e-commerce backend

An e-commerce backend on Java and Spring Boot, with Kafka for events, Temporal for durable workflows and PostgreSQL for storage, deployed to AWS EKS. Phase 2 is scaffolded as a Gradle monorepo.

  • Java
  • Spring Boot
  • Kafka
  • Temporal
  • PostgreSQL
  • EKS
More from the studio
  • Anchor · scam defense
  • Natural World in Numbers · measurement
  • LabelCheck · label compliance
  • Farm AI · drone analytics
50 Limes case study →

Earlier work

University of Washington · Informatics
Clique

Capstone. A platform that connects students with trusted friends and family to build a financial community that tackles student debt.

View ↗
Chores

Helps families share household chores, with points and rewards for getting them done.

View ↗
Music trends visualization

A React app that visualizes music trends and greatest hits from public data.

View ↗
Image Tracing

Trace any image live and download the result. Built as a tool for practicing drawing.

View ↗
Incarceration in Washington

Analysis of incarceration in Washington State from 2004 to 2013, broken down by race.

View ↗
US mass shootings, 2016

A map of 2016 mass-shooting data showing which cities had the most and fewest incidents.

View ↗
Human trafficking analysis

A historical look at human trafficking up to the present day.

View ↗
// 04

Education & recognition

Education

M.S. Computer Science, Data Science

Spring 2027

Seattle University · Seattle, WA · in progress

Graduate Certificate, Computer Science

Seattle University · Seattle, WA

B.S. Informatics

University of Washington, Information School · Seattle, WA

Recognition & credentials

Microsoft Patent Award

"Data Flow Control in Distributed Computing System"

CORe Credential of Readiness

Summer 2023

Harvard Business School Online · business analytics, economics, financial accounting

The New Technologies (TNT) Mentor

Summer 2023

Microsoft

// 05

Skills

Languages
  • Python
  • C#
  • Java
  • Go
  • C++
  • TypeScript
  • JavaScript
  • SQL
  • KQL
  • SCOPE
Frameworks & tools
  • .NET
  • React
  • Angular
  • Spring Boot
  • Bot Builder SDK
  • Docker
  • Git
  • Selenium
Cloud & data
  • Azure Data Explorer
  • Azure Data Factory
  • Azure OpenAI
  • Cosmos
  • AWS (EC2, EKS)
  • Kafka
  • PostgreSQL
  • MySQL
Expertise
  • Distributed systems
  • API development
  • Big-data pipelines
  • Rate limiting
  • Reliability & on-call
  • CI/CD
// 06

Arts

Away from the keyboard I paint in oil, mostly Pacific Northwest landscapes, and I take photographs on the ground and from the air with a drone. I also build furniture.

Oil painting

Pacific Northwest landscapes.

Photography

On the ground and in the air.

Aerial

Drone photography and video. I'm an experienced multirotor pilot.

Carpentry

Furniture, built by hand.
// 07

Contact

Email is the fastest way to reach me.

I'm glad to talk about distributed systems, reliability, AI infrastructure, or technology for Ethiopia and its diaspora.

Emailkidussendeke@gmail.com→ LinkedInlinkedin.com/in/kidus-sendeke↗ GitHubgithub.com/Kidusyohanes↗ Resumeresume.pdf↗