Sieve Tech Stack

Video datasets and infrastructure for frontier AI models

Software Development San Francisco, CA 11–50 employees Founded 2022 Privately Held

Sieve is building video training data at scale for large AI labs and Fortune 100 companies. The tech stack—PyTorch, Go, Rust, GCP, AWS, Kubernetes (Argo CD, Helm, Kustomize)—reflects a systems-heavy engineering org optimizing cost and throughput for petabyte-scale video processing. Active projects center on ML+ETL pipeline orchestration and video understanding, with hiring accelerating across engineering roles (6 open positions), signaling the company is scaling data infrastructure faster than data science or go-to-market.

Tech Stack 25 technologies

Core StackPython PyTorch GitHub Go Rust TypeScript Cursor React Next.js React Native Svelte Terraform AWS Oracle Cloudflare Prometheus OpenTelemetry Helm C++ CI/CD Claude GCP VictoriaMetrics Argo CD Kustomize

What Sieve Is Building

◆Challenges

High-quality training data bottleneck
Cost-effective processing of petabytes of video data
Optimizing compute scheduling for large video pipelines
Cost-effective ml pipelines
Optimizing system uptime
Handling large-scale video data
Scaling data pipelines
Ensuring high-quality data delivery
Improving system efficiency
Scaling high-growth startup

▲Active Projects

Building internal tooling and ci/cd for rapid iteration
Video understanding pipelines
Ml+etl pipeline orchestration for large video data
Ml + etl pipeline orchestration
Data pipeline development
Ml filter development
Internal qa dashboard
Build recruiting function
Host recruiting events
Video collection platform

Hiring Activity

Accelerating10 roles · 8 in 30d

Department

Engineering

Data

Ops

Product

Sales

Seniority

Mid

Lead

Senior

Junior

Company intelligence

Find more companies like Sieve by tech stack, pain points and active projects

Get started free

About Sieve

Sieve operates an AI research lab focused exclusively on video data, assembling exabyte-scale infrastructure, video understanding techniques, and diverse data sources into training datasets for video modeling. The company serves frontier AI research labs, Fortune 100 enterprises, and generative AI startups. Founded in 2022 and based in San Francisco, Sieve is 11–50 employees with recent hiring velocity focused on engineering. Core operational challenges center on cost-effective processing of petabyte-scale video, optimizing compute scheduling, and scaling data pipelines while maintaining quality delivery.

HeadquartersSan Francisco, CA

Company Size11–50 employees

Founded2022

Hiring MarketsUnited States

Frequently Asked Questions

What tech stack does Sieve use?

Core stack: Python, PyTorch, Go, Rust, C++. Infrastructure: GCP, AWS, Kubernetes (Argo CD, Helm, Kustomize), Terraform, Cloudflare. Observability: Prometheus, OpenTelemetry, VictoriaMetrics. Frontend: React, Next.js, TypeScript.

What is Sieve working on?

Active projects include ML+ETL pipeline orchestration for large video data, video understanding pipelines, internal CI/CD tooling, ML filter development, and a video collection platform. Also building out recruiting function.