Alectio
Machine Learning-Driven Data Curation, DataPrepOps platform
Quick facts
- Company
- Alectio
- Service type
- Data annotation / AI training data
- Specialties
- Text
- Hiring status
- Both: hires workers and takes vendor projects
- Website
- http://www.alectio.com
- Careers
- Unavailable
- Profile last verified
- 2026-01-29
Application process overview
Alectio was a Santa Clara-based DataPrepOps startup founded in 2019 by Jennifer Prendki, focused on active-learning tools that reduced the amount of labeled data needed to train models. It employed annotators and ML engineers in a traditional startup model rather than running a crowd platform, and appears to be effectively defunct in 2026.
Key findings
Application Process: Historically standard startup hiring via LinkedIn/Glassdoor postings. There is no open contributor signup - Alectio was not a gig platform.<br><br>Assessments: Role-specific technical interviews (ML engineering, data ops) rather than annotator tests.<br><br>Job Types / Expertise: Primarily software engineering, ML, and product roles; some internal labeling staff. Not a destination for independent annotators.<br><br>Compensation: Glassdoor shows a compensation-and-benefits rating of 2.0/5 with reports of missed monthly salary payments.<br><br>Flexibility: Previously fully remote with flexible hours - one of the few positives cited by former employees.<br><br>Challenges / Concerns: Multiple Glassdoor reviews criticize the CEO for erratic management, unrealistic deadlines, and alleged non-payment. The original domain alectio.com now serves unrelated gambling content and the founder's LinkedIn lists a 'stealth post-LLM' startup, strongly suggesting Alectio is wound down.<br><br>Legitimacy: Was a legitimate VC-backed startup (CB Insights-listed) but current operational status is doubtful.
Conclusion
Alectio is best treated as defunct or dormant as of 2026. For anyone researching it as a job option, the combination of a hijacked-looking domain, founder pivot, and strongly negative employee reviews make it a poor choice. It remains relevant only as a historical reference in the DataPrepOps/active-learning space.