Data Engineers: $70-$190/hr from Frontier AI Labs

Open application· 79 spots this round

$70-$190/hr data engineering work, on your schedule

Review AI-generated pipelines, transformations, and warehouse models the way you'd review a PR before it hits the lake. Catch the silent data loss, the broken idempotency, the model that fans out wrong. The judgment that keeps data trustworthy at scale is exactly what AI labs need.

Fully remoteYour scheduleWeekly pay

Apply nowApply once, get matched on a rolling basis. No prior AI experience needed.

Trusted by top research companies

Rated by experts worldwide

Hi, we're Zac and Jack, the founders of Terac. We want to talk to you directly, because you are the most important part of what we're building.

Terac is a community of experts. People who have spent years getting good at something specific and hard. The world is about to need more of you, not less. As AI takes on more of the world's work, the bottleneck shifts to the people who actually know what they're talking about.

Expert labor is the rarest resource in the world right now, and it is shockingly hard to find. The companies that need a data engineer's eye on a pipeline that silently drops rows spend weeks chasing people, paying placement fees, and settling for whoever is available. Meanwhile thousands of qualified people are sitting with knowledge that no one ever asks for.

That gap is what we're here to close. Every project that lands on Terac is routed to the people who actually know the answer, on their schedule, paid fairly, and only when the work is verified. No middleman taking a cut of your time. No vague gigs. No chasing checks.

We care about every single person in this community. If you join Terac, you're not a row in a database to us. We read the feedback. We answer the emails. We will fight for you when a customer is being unreasonable, and we will be honest with you when something on our side is broken. The quality of this panel is our entire company, and we owe you a serious bar.

If you've made it this far, here is what we're asking: claim your profile. Put your expertise on the record. Let the world's most ambitious teams come find you for the work only you can do.

Zac & Jack

Founders

Data Engineering questions

Still curious? Write to us at support@terac.com.

Narrow specializations are exactly what is most valuable here. Frontier models struggle most with opinionated, tool-specific reasoning, so a deep dbt practitioner who can explain incremental materialization tradeoffs, or an Iceberg expert who can spot a poorly chosen partition spec, produces higher-signal training data than a generalist. You will be matched to tasks that fit your actual stack, not asked to evaluate work outside it.

Tasks vary by project, but common deliverables include reviewing AI-generated SQL transformations and dbt models for correctness and style, evaluating orchestration code written in Airflow or Prefect, assessing data quality rule definitions, and writing worked examples that show how you would design a schema or debug a pipeline failure. You will see a task description before committing, so you can decline anything outside your comfort zone.

Certifications like the Databricks Certified Data Engineer tiers or the dbt Analytics Engineer certification are used as one signal during profile verification to calibrate your seniority and tool familiarity, but they are not a hard gate. Equivalent demonstrable experience on those platforms counts just as much. Where they do help is in routing you toward tasks that specifically involve Spark, Delta Lake, or lakehouse patterns, since reviewers with that background produce more reliable evaluations on those topics.

No task will require you to use, reference, or reproduce proprietary assets from your employer. All evaluation scenarios are either fully synthetic or based on anonymized, publicly available schemas and datasets. Your job is to assess the AI's reasoning and correctness using your own expertise, not to expose or replicate anything from your current role.

It makes you more suitable for a specific subset of work. Tasks that ask models to reason about data lineage requirements, audit logging, PII handling strategies, or access control design in regulated environments require reviewers who actually understand what correct looks like in those contexts. Your HIPAA or SOX experience will be noted on your profile and used to route you toward evaluations where that background is directly relevant.

Why your expertise matters

Data engineering sits at the foundation of every ML pipeline, and AI models consistently produce plausible-looking but subtly broken code for tools like Apache Spark, dbt, and Airflow that only a practicing engineer would catch. Errors in schema design, partition strategy, or incremental load logic can silently corrupt downstream models in ways that are invisible to generalist reviewers. Your ability to recognize whether a pipeline will actually perform at scale, handle schema drift gracefully, or violate a platform's concurrency limits is exactly the judgment that makes AI-generated data infrastructure trustworthy.

How pay works

Rates toward the $190/hr ceiling go to engineers with demonstrated depth in a high-demand combination: cloud-native warehousing on Snowflake, BigQuery, or Redshift alongside orchestration experience in Airflow or Dagster, or specialists in streaming architectures using Kafka and Flink. Work is fully remote, billed by the verified hour, and payment releases only after your deliverable passes Terac's completion check, so there are no unpaid hours and no ambiguous project-end disputes.

What the work looks like

A sample of the data engineering work you would pick up. Every project is scoped, remote, and paid on verified completion.

Review an AI-generated dbt model for correct incremental materialization logic and flag grain mismatches that would cause silent fan-out in downstream aggregations.
Evaluate a Spark job written by an AI to identify shuffle-heavy joins, missing partition pruning, and executor memory settings that would fail on a production-scale dataset.
Assess an Airflow DAG proposed by an AI for improper dependency chaining, missing SLAs, and catchup behavior that could cause duplicate loads in an idempotent pipeline.
Write a worked example of a correct CDC ingestion pattern using Debezium and Kafka into an Iceberg table, annotated to explain why each design decision avoids data loss during schema evolution.
Score AI-generated SQL for a Snowflake warehouse query against criteria including clustering key alignment, use of transient tables, and unnecessary cross-database joins.
Identify whether an AI-proposed data vault implementation correctly separates hubs, links, and satellites and whether the business key selection would survive a source system migration.