Howdy, I'm Ike. I am a machine learning engineer working at GetReal Security where I spend a lot of time talking about and working on the machine learning lifecycle. I joined GetReal to focus on scaling ML serving for our live-meeting product GetReal Protect, which mitigates real-time deepfake threats across all major video conferencing platforms. I get to work with my favorite languages, Python and Rust, and spend time scaling my favorite database, Postgres.
Previously, I spent time consulting for an AI startup, operating as their founding engineer, where I got to work on hosting and scaling LLM inference in on-prem and air-gapped environments.
Before that, I had five years working horizontally as a data scientist and data engineer at a consulting firm. My time there yielded some great "developer lore":
- Deploying a database on an FPGA that got sent to the moon on a lunar lander that subsequently fell on its side
- Collaborting on a proof-of-concept algorithmic anti-cheat software that is now used in a video game franchise enjoyed by over 100 million monthly active users
- Modeling transmission of COVID-19 for the largest prison system in the world during a global pandemic
I like to approach software and systems from first principles, breaking things down to their component parts in order to build a better understanding of the instrinsics. Typically the software I like to tear down to the studs are things like databases, distributed systems, and machine learning; though I am known to churn through other topics on a whim. Additionally, I care a lot of my developer experience. I am guilty of staying up far too late working on ricing my terminal experience.
I have an academic background in physics and nuclear engineering. During my studies I was able to research with some incredible colleagues at the Nuclear Engineering Teaching Laboratory, which hosts a functional research reactor. Though my time was largely spent writing Monte Carlo N-Particle Transport code, a flavor of Fortran 90. And since nobody asked, my favorite equation is the diffusion equation.
$$\dfrac{\partial\phi(\boldsymbol{r}, t)}{\partial t} = \nabla \cdot [D(\phi, \boldsymbol{r})\nabla\phi(\boldsymbol{r}, t) ]$$