About Streamforge

Built by engineers who've been paged at 3 a.m.

In 2023, Karen Sullivan left her role managing a 200-topic Kafka pipeline at a Minneapolis-based freight logistics company and started Streamforge. She'd watched schema drift cause three separate downstream audit failures in two years. Each time, the fix was hours of offset forensics and producer-team Slack threads. There was no tool that caught the problem before it propagated. She decided to build one.

Our mission: make silent pipeline failures extinct

Real-time event pipelines are load-bearing infrastructure. A corrupt event stream silently produces wrong analytics, wrong invoices, wrong inventory counts. Consumer lag metrics don't catch it. Distributed tracing doesn't catch it. The failure mode lives in the schema contract between producers and consumers, and that contract has no enforcer.

Streamforge's job is to be that enforcer. Catch every deviation before it propagates. Not after a customer notices. Not after an audit. Before.

The team

Karen Sullivan, CEO and Founder of Streamforge

Karen Sullivan

CEO & Founder

Managed Kafka infrastructure for a freight logistics company from 2020 to 2023, growing the pipeline from 40 to 200+ topics. Ran three post-mortems for silent data loss events that existing tools never flagged. Founded Streamforge in 2023 to build what she kept wishing existed.

Marcus Okafor, Co-Founder and CTO of Streamforge

Marcus Okafor

Co-Founder & CTO

Spent four years at a fintech building Flink connectors for real-time fraud-signal pipelines. Focused on exactly-once semantics and checkpoint-based recovery before joining Karen in 2023 to architect Streamforge's self-healing routing engine.

Anya Petrov, Head of Engineering at Streamforge

Anya Petrov

Head of Engineering

Led data platform engineering at a healthcare analytics company, where a silent Kinesis event-loss event caused a week of stale patient-encounter aggregates before the data team noticed. Joined Streamforge in 2024 to own the schema monitoring engine and DLQ Inspector internals.

How we build

Correctness over features

We'd rather do schema validation right than ship six half-baked detection modes. Our baseline accuracy matters more than our feature count.

No false positives

An alert that fires when nothing's wrong trains teams to ignore alerts. We obsess over precision. Alert fatigue is a product failure, not an operator problem.

Operator trust

We never write to your broker, never store your event payloads. You're trusting us with infrastructure credentials. We take that seriously, full stop.

Minneapolis, Minnesota

Minneapolis was the right call. Lower burn rate, direct-to-engineer culture, and no one here cares about your deck. The midwest industrial pragmatism is in the product: we'd rather be precise than promotional.

Founded 2023. Bootstrapped. We're not funded by a VC who needs to show a 10x return in three years. We grow when our customers' pipelines get more reliable. That alignment matters to us.

80 South Eighth Street, Suite 900
Minneapolis, MN 55402
[email protected]
Company snapshot
Founded 2023
Stage Bootstrapped
Team size 3 engineers
Focus Pipeline integrity