Hire reliable workers from a pipeline you can trust.
Astris HR connects you with refugee, immigrant, and second-chance candidates pre-screened, self-validated, and ranked by a model that understands transferable skills. The pipeline UI is yours.
Upload the JD. Astris does the rest.
Upload your JD
PDF, DOCX, or paste. Astris HR Model parses skills, certifications, responsibilities, schedule, benefits, languages, and physical requirements.
Review and edit
Edit any field Astris got wrong, rate the parse, and the model auto-corrects. Publish when you're ready.
See ranked candidates
Each match shows the 7-component score breakdown, transferable-skill bridges, gaps, AI rationale, and retention forecast.
Run the pipeline
9-column kanban: Potential → Screening → R1 → R2 → R3 → Offer → Hired → Active 30+d → Left/Rejected. Status moves trigger emails.
Original resume. Astris version. Side by side.
Every candidate card shows their original resume alongside the structured Astris version, the 7-component score breakdown, gap badges, transferable-skill bridges, and a plain-language rationale. Contact via email goes through Astris and logs to the placement timeline.
- 7-component score with per-dimension breakdown
- Transferable-skill bridges shown with from → to
- Gap badges flag missing certifications and constraints
- Plain-language rationale via Astris HR Model
- One-click status moves through the 9-column pipeline
- One-click 'Contact via Astris' email — logged on the timeline
- Crop Cultivation → Inventory Handling
- Tractor Operation → Forklift Operation
- Team Leadership → Team Leadership (direct)
What employers ask, answered.
How do candidates get on the platform?
Caseworkers at refugee resettlement and workforce development organizations upload resumes (or generate them from a conversation). Astris parses, candidate self-validates, profile enters matching.
Can we trust the resume data?
Yes. Every candidate reviews their Astris-parsed profile and approves it before any employer can see it. If the parse is wrong, the candidate rejects and the caseworker corrects.
What languages do you support?
Resume parsing in 35+ languages, UI in 8, follow-up SMS in 8. Including Pashto, Dari, Tigrinya, Karen, Rohingya, Sorani Kurdish, and Levantine Arabic.
How is the score computed?
Seven components — skills, experience, language, distance, schedule, certifications, transportation — weighted 40/20/10/10/10/5/5. Composite is a weighted average with a transferable-skill graph boost and a company-quality adjustment.
What about retention?
Each match comes with retention probabilities at 30, 90, 180, and 365 days. After hire, multilingual SMS follow-ups fire at those milestones. Retention outcomes feed the next model retraining.
Do you integrate with our ATS?
Not yet. REST API exposing matches + placements is on the roadmap. For now, the platform stands alone for refugee/immigrant hiring workflows.