Anshul
Shanker
Engineer at heart. I build systems, think about how teams build software, and when something's worth it, I write it down.
“Simplicity is a prerequisite for reliability.”— Edsger Dijkstra
I'm an engineering leader and software architect in Chicago. At OneMain Financial I own architecture across three product domains, including the loan-servicing experience over a million customers use every month — a stack of distributed services that really can't have a bad day.
I take complex, large-scale initiatives from roadmap to delivery — reducing hard problems to simple solutions and driving them to production. I care about systems built for predictability and resilience, and about owning both the architecture and the execution that gets them there.
That same instinct — owning the whole problem — is where PACE came from. As AI became part of how engineering teams work, a pattern emerged: individual velocity up, but design decisions getting made implicitly inside AI sessions, with no shared context and no review gate before code was written. The result was drift — patterns diverging, intent lost, debt accumulating faster than anyone could track. PACE is the framework I introduced to fix that — and its real advantage isn't the model, it's memory: every approved design becomes context the next one builds on, so the system gets sharper the more the team uses it.
📍 Chicago, IL
PACE: Plan, Agentic Context, Execute
A framework for safe AI adoption in engineering teams. Its real advantage isn't model capability — it's compounding organizational memory: a living context layer of structured decisions and versioned blueprints, a human-gated review before any agent writes application code, and every approved design making the next one better.
Read the paper →Say hi
Always happy to talk systems, scale, or how teams ship with AI.