Tanner Yohe

Principal Software Engineer

Tanner Yohe

I build and run distributed systems, and the internal platforms other engineers ship on.

building distributed systems
8 yrs
credit applications a day on the Temporal platform I led
60K+
compute spend cut, with zero downtime
50%
to stand up a new service, down from a week
½ day

About

Hi, I'm Tanner.

I'm a software engineer with eight years of building and operating distributed systems in Python, Ruby, and C#/.NET. I'm currently a Principal Software Engineer at Avant, where I led an 8-month re-platform of the lending core onto Temporal: 100+ workflows carrying all origination traffic, shipped with zero downtime and a 50% cut in compute spend.

I like building the platforms other engineers ship on, and I own what I build in production, on-call included. I do my best work on lean, high-autonomy teams where engineers own problems end to end and move fast without giving up quality.

Experience

Where I've worked

  1. Avant

    Chicago, IL

    • Principal Software EngineerJan 2026 – Present
    • Senior Software EngineerJan 2025 – Jan 2026
    • Led the re-platform of the core lending monolith onto Temporal, replacing an implicit workflow model with 100+ explicit workflows, including parent workflows that orchestrate nested children. It carries all origination traffic: 60K+ credit applications a day, each fanning out to tens or hundreds of Temporal API calls.
    • Rolled the migration out incrementally over 8 months across 20+ engineers in four pods. Shipped with zero downtime, cut stuck credit applications from about 100 a day to under 1, and reduced compute spend 50%.
    • Built the internal service platform for a 50-engineer organization: CI/CD workflows, database access and migration patterns, REST and gRPC controller blueprints, and service-layer structure. Adopted by nearly half the company's Python microservices; new service setup dropped from about a week to half a day.
    • Led a tiger team through a major rewrite inside the core monolith, unblocking features that drove a 2x increase in credit issuance.
    • Own on-call for the originations division (8 services, including the core monolith) at under one minute mean time to response and under one hour to resolution.
    • Present architecture and technical direction to engineering leadership, feature teams, and business stakeholders.
    • Python
    • Litestar
    • Temporal
    • gRPC
    • Kafka
    • PostgreSQL
    • DynamoDB
    • Redis
    • AWS
    • Kubernetes
    • Ruby
    • Rails

  2. Recorded Future

    Boston, MA

    • Team LeadApr 2024 – Jan 2025
    • Data EngineerOct 2022 – Apr 2024
    • Designed and built about 25 Python microservices on AWS within a 300+ service data platform; the busiest peaked at 15 TB ingested per day.
    • Set data engineering standards across a 30-engineer division: moved relational workloads off MongoDB onto PostgreSQL and Aurora, introduced a standard project layout and toolchain, and split extract, transform, and load into independently deployable services.
    • Led two pods totaling 10 engineers and carried on-call across the 300-service estate at under 30 minutes mean time to resolution.
    • Python
    • PostgreSQL
    • Aurora
    • MongoDB
    • Elasticsearch
    • RabbitMQ
    • AWS
    • Kubernetes

  3. Torch Technologies

    Huntsville, AL

    • Software Design EngineerDec 2018 – Oct 2022
    • Built and maintained two C# .NET web APIs on Entity Framework serving the core training platform and partner LMS integrations.
    • Designed a Python and C# ETL pipeline that ingested dirty, incomplete data from a proprietary third-party format into PostgreSQL.
    • Trained engineers on internal systems, mentored recent graduates, and maintained containerized Jenkins CI/CD pipelines across projects.
    • Developed training apps in Unreal Engine, plus custom launchers and data entry portals in React.
    • C#
    • .NET
    • Entity Framework
    • PostgreSQL
    • Python
    • C++
    • Jenkins
    • Docker
    • Unreal Engine
    • React

  4. Publicis Sapient

    Chicago, IL

    • Junior Associate, TechnologyJun 2018 – Nov 2018
    • Redesigned internal and external systems for Fortune 500 clients across product information, point of sale, and microservice architectures.
    • Java
    • MongoDB
    • SQL
    • Angular
    • React

  5. Yin Lab, Northern Illinois University

    DeKalb, IL

    • Research AssistantMay 2017 – May 2018
    • Built highly cited software tools and high-traffic web servers supporting peer-reviewed molecular genetics research.
    • Python
    • Flask
    • SQL

Projects

Things I build for fun

Race Rewind

In progress

Replay any NASCAR race lap by lap on a map of the track.

Pick a season, series, and race, then watch the field run on the real track layout with the running order, gaps, pit stops, penalties, and a race feed. Playback runs from 0.25× to 64×, with a scrubber marked at cautions, stage ends, and penalties.

  • Reconstructs every car’s position from lap times the way timing and scoring does, so the running order changes at the line.
  • Animates pit stops from pit data: braking onto pit road, driving to the box, and accelerating back to speed.
  • Infers when penalties are served from lap notes when the official infraction records are empty.
  • TypeScript
  • React
  • Vite
  • Tailwind
  • TanStack Query
  • Python
  • Litestar
  • Aurora DSQL

nascar-api

A typed Python client for NASCAR's undocumented Content Feed API.

Historical and live race data across the Cup, Xfinity, and Truck series: schedules, results, lap times, lap notes, pit stops, and loop statistics. Published on PyPI.

  • Fully typed models built on Pydantic for schema validation.
  • Retry policy built on tenacity that retries 5xx responses and decode failures and skips 403s.
  • Python
  • Pydantic
  • tenacity

Rally

A desktop dashboard for the GitHub PRs you review, with Claude drafting the first pass.

Rally tracks the pull requests waiting on you, sorts them into review phases, and runs a Claude Code review skill of your choice to draft comments. You edit the drafts, refine them with follow-up turns, and submit to GitHub without leaving the app.

  • Works out each PR’s review phase from GitHub state, including force-pushes and re-requested reviews.
  • Runs reviews as resumable Claude Agent SDK sessions that hand drafts back through an in-process MCP tool.
  • Posts reviews directly through the GitHub API, folding any comment GitHub rejects into the review body.
  • Ships macOS and Windows builds with semantic-release on every push to main.
  • TypeScript
  • Electron
  • React
  • Tailwind
  • Claude Agent SDK
  • GitHub API

More on GitHub.

Skills & interests

What I work with, and what I'm into

Languages

  • Python
  • Ruby
  • C# / .NET
  • SQL
  • TypeScript

Systems

  • Temporal
  • Kafka
  • gRPC
  • REST
  • Litestar
  • Rails
  • PostgreSQL
  • Redis

Platform

  • AWS
  • Kubernetes
  • Docker
  • CI/CD
  • On-call & incident response

Also worked with

  • DynamoDB
  • Aurora
  • MongoDB
  • Elasticsearch
  • RabbitMQ
  • Entity Framework
  • React
  • Java
  • C++
  • Unreal Engine

  • Durable workflows

    Making long-running business processes explicit, observable, and recoverable. Temporal is my current tool of choice.

  • Developer platforms

    Paved roads, service blueprints, and tooling that shrink the time from idea to production for a whole organization.

  • Motorsport data

    Turning NASCAR's timing and scoring feeds into tools that show how a race actually unfolded.

  • AI-assisted engineering

    Putting coding agents to work on the tedious parts of review and delivery, with an engineer making the calls.

  • Machine learning

    The focus of my master’s at Georgia Tech, and a toolbox I keep coming back to.

Contact

Let's talk

Whether it's a role, a hard distributed systems problem, or NASCAR data, I'm happy to hear from you. Email is the fastest way to reach me.