Remote Data Jobs

Roughly: data analysts answer business questions with SQL, dashboards, and analytics tools. Data scientists do that plus statistical modeling, experimentation, and sometimes light ML. Data engineers build the pipelines, warehouses, and infrastructure that everyone else relies on. ML engineers ship machine-learning systems into production.

Companies inconsistently use these titles. A 'data scientist' at one company is a 'senior analyst' at another and a 'product analyst' at a third. Read the actual responsibilities, not the title, and ask in interviews what fraction of the day is SQL versus modeling versus infrastructure.

Data hiring is heavier on take-homes than engineering. Expect: recruiter screen, SQL screen (live or take-home), case study or business problem walkthrough, technical depth interview, and cross-functional partner panels. Total time is 3-6 weeks.

For data scientists and ML engineers, expect probing on statistical reasoning, experimentation pitfalls, and modeling trade-offs. For data engineers, expect SQL + Python + system-design depth.

Data work is more naturally async than most engineering work because the deliverables (queries, models, dashboards, written analyses) are durable artifacts. Overlap requirements are typically 3-4 hours with the primary product team you support.

Beware roles where 'data scientist' translates to 'on-call dashboard requester' — they require constant real-time availability and are exhausting to do remotely across timezones.

Data pay tracks engineering closely. Data engineers and ML engineers typically earn at parity with software engineers at the same level. Data scientists and senior analysts are usually slightly below software engineers at the same level. ML research and applied research roles can exceed standard engineering compensation at specific companies.

Cross-border, expect the same EOR + tier-by-region patterns as engineering. Specialized ML and inference-infrastructure roles increasingly carry premium pay because supply is tight.

Full-time embedded roles dominate. Fractional data leadership (head of data on a 1-2-day-a-week basis) is common at very early-stage startups. Pure-contract analyst work exists but is competitive and lower-paid than full-time.

Data-specific concerns: unpaid 'take-home' assignments that ask you to analyze the company's actual production dataset (you are doing free consulting), 'ML research' fronts that want you to produce models for resale without IP terms, and 'crypto data' roles that turn out to be wash-trading or market-manipulation infrastructure. Read the company's product carefully and walk away if the use of your work is ambiguous.

AI has compressed the time required for routine analysis and dashboard generation but has expanded demand for: people who can build trustworthy evaluation pipelines for LLM systems, ML engineers who can ship production AI features, and data engineers who can manage the cost and reliability of AI workloads.

Analyst roles weighted toward ad-hoc dashboarding have seen the most pressure. Roles weighted toward causal inference, experimentation, and applied modeling have grown.