Up-Skill
An AI career assistant that scores resumes ATS-style, runs mock interviews, maps skill gaps, and proposes personalized learning paths.
Up-Skill is a career assistant that turns a candidate's resume and target role into a structured development plan. It scores the resume ATS-style, runs a conversational mock interview, identifies skill gaps against the role, and proposes a personalized learning path.
Career tooling is fragmented: resume scoring, interview prep, and learning plans live in separate products. Candidates rarely get a single coherent view of where they stand and what to do next.
Build a multi-stage LLM workflow where each stage produces structured output the next stage consumes. Use Groq for low-latency reasoning and Mistral for nuanced evaluation. Stitch and Supabase persist profile state and learning progress.
Flask service · resume parser · ATS scoring stage · mock interview (conversational) stage · skill-gap analysis stage · learning-path synthesis stage. Profile state and progress persisted in Supabase. Stitch used for design system.
Technical Challenges
Stage-to-stage consistency was hard — an ATS score from one model had to be interpretable by the skill-gap stage in another. A shared structured schema between stages resolved most drift.
Engineering Learnings
Multi-model pipelines need a strong contract between stages. Once outputs were schema-bound, swapping models became a tuning decision instead of a rewrite.