AstraWork
An AI-powered talent matching system that understands candidates through deep evidence (Resume, GitHub, Projects, Experience) instead of relying only on basic keyword matching.
THE BUSINESS PROBLEM
Recruitment still depends heavily on resumes and keywords.
The problem is simple: keywords show what someone says they know, not what they can actually do.
Good candidates get filtered out because their experience does not match the wording of a job description.
THE IDEA
AstraWork looks beyond the resume.
It combines experience, projects, GitHub and skills to match candidates with roles based on evidence and capability, not just keywords.
THE BUSINESS & PROCESS ANALYSIS
I mapped the recruitment journey from job requirement → candidate profile → screening → shortlist.
The biggest gap was the screening stage, where most candidate context gets lost.
The solution focused on bringing that context back into the decision.
THE REQUIREMENTS & IMPLEMENTATION
Core requirements included candidate profiling, evidence extraction, role analysis and explainable candidate matching.
AI was used to interpret unstructured information and turn it into useful matching signals.
THE RESULT / CURRENT STATUS
AstraWork reached the prototype stage.
The project helped validate one idea: better hiring may start with better understanding, not more applications.
WHAT I LEARNED
The hard part wasn't the AI.
It was defining what counts as evidence and how that evidence should influence a hiring decision.