Enterprise to Startup: Lessons from Building at Both Scales
What I learned switching between enterprise systems at Ameriprise and building a startup from scratch with GLINCKER.
Gagan Deep Singh
Founder | GLINR Studios
For the past few years, I've been living a dual life. By day, building enterprise financial systems. By night (and weekends), building a startup from scratch. The contrast has taught me more about software engineering than either experience alone.
Enterprise: The Art of Constraints
At Ameriprise Financial, I learned to work within constraints that most developers never encounter:
- Regulatory compliance — every change goes through security review
- Legacy systems — integrating with COBOL services from the 1990s
- Scale — serving millions of financial transactions daily
- Team coordination — dozens of teams, hundreds of services
The biggest lesson? Boring technology is reliable technology. When millions of dollars flow through your system, you don't want "bleeding edge." You want battle-tested.
Startup: The Art of Speed
GLINCKER is the opposite. As the founder and lead architect, I make every technical decision. The constraints are different:
- Ship fast — users don't wait for perfect
- Budget — every AWS dollar matters
- Wear every hat — architect, developer, DevOps, support
- Iterate — what you build today might get thrown away tomorrow
The biggest lesson here? Perfect is the enemy of good. A shipped feature beats a planned masterpiece every time.
What I Took From Each
From enterprise, I took:
- Architecture discipline — design docs, ADRs, and proper planning
- Testing rigor — unit, integration, and E2E testing as a non-negotiable
- Observability — logging, metrics, and tracing from day one
From startup, I took:
- Pragmatism — solve the problem, not the abstraction
- User empathy — talk to users before writing code
- Speed — CI/CD, feature flags, and rapid deployment
The Sweet Spot
The best code I write lives at the intersection. Enterprise architecture with startup velocity. Proper design patterns without over-engineering. Tests where they matter, not everywhere.
This is what I bring to every project now — whether it's building at Marriott or shipping the next GLINCKER feature.