Reflection AI

Reflection AI · New York, NY +1 · Onsite · Sep 18

Architect security protecting multi-million-dollar GPU training capacity.

Reflection AI · San Francisco, CA +2 · Onsite · Jul 21

Lead ingestion layer turning open web and licensed corpora into training datasets.

Reflection AI · New York, NY +1 · Onsite · Oct 9

Build compliance programs across AI governance, data, trade, privacy and contracting.

Reflection AI · New York, NY +2 · Onsite · Oct 1

Build shared services powering research, training, and production environments.

Reflection AI · San Francisco, CA +2 · Onsite · Sep 30

Build the trusted data backbone for open-weight AI systems.

Reflection AI · New York, NY +2 · Onsite · Sep 30

Build core data systems powering research, training, and production environments.

Reflection AI · New York, NY +1 · Onsite · Jul 14

Architect security for multi-million-dollar GPU clusters protecting model weights.

Reflection AI · New York, NY +2 · Onsite · Mar 12

Build and operate the foundational platform every engineering team depends on.

Reflection AI · San Francisco, CA +2 · Onsite · Mar 24

Build and scale systems powering frontier model pre-training.

Reflection AI · San Francisco, CA +2 · Onsite · Jul 15

Lead the foundational platform every engineering team depends on daily.

Reflection AI · New York, NY +2 · Onsite · Apr 30

Embed with pre-training ML and Data teams at frontier model core.

Reflection AI · San Francisco, CA +2 · Onsite · Oct 2025

Research scaling laws and data efficiency for large language models.

Reflection AI · New York, NY +2 · Onsite · Mar 12

Design core infrastructure powering frontier open AI models at massive scale.

Reflection AI · Seoul · Onsite · Jun 9

Own end-to-end delivery of agentic apps from pre-sales discovery to production.

Reflection AI · New York, NY +2 · Onsite · Mar 20

Keep large-scale GPU compute layer healthy and highly available.

Reflection AI · New York, NY +1 · Onsite · Dec 2025

Shape model performance on agentic tool use and long-horizon reasoning.