YesResume
View Live ProjectBusiness Context
Most candidates reuse generic resumes and manually adjust them per role, often missing critical ATS keywords and formatting constraints.
YesResume automates this by
- Extracting structured requirements from job descriptions
- Prioritizing relevant experience based on role alignment
- Optimizing keyword coverage for ATS systems
- Enforcing strict one-page formatting constraints
Engineering Architecture
Dual Generation Modes
Direct mode (30–60s) for fast turnaround, plus an adaptive interview mode for deeper, higher-quality optimization.
Multi-LLM Orchestration
Supports OpenAI, Gemini, and Cerebras with task-based routing and fallback logic. The multi-model design enables experimentation and optimization across cost, speed, and output quality by leveraging each model's strengths.
Deterministic LaTeX Engine
AI generates structured content while multiple deterministic LaTeX templates enforce layout consistency and page-fit control.
Schema-Controlled Output
JSON schema validation and strict TypeScript typing ensure consistent, production-safe outputs before PDF compilation.
Key Trade-offs
Creativity vs. Control
Separated AI content generation from deterministic layout enforcement to maintain structural reliability.
Architectural Complexity vs. Model Agility
Introduced multi-model orchestration to experiment and optimize cost, speed, and quality, accepting increased system complexity for greater adaptability and long-term flexibility.
Speed vs. Depth
Two generation paths balance immediate results with more refined, interview-driven tailoring.