Sage

The description states that Sage:

• Parses a user's resume in-browser (using pdf.js and mammoth)
• Uses GPT-5.6 with structured outputs via JSON schema
• Compares resume content against job postings to determine fit scores
• Provides both individual posting analysis and role-type shortlist analysis
• Offers "recurring gaps" as coaching directions, not project specs
• Never stores user data on servers; all processing happens client-side except model calls

Inference Sage appears to be a prototype or proof-of-concept built for a hackathon. It is not evidenced to have any production deployment or customer base.

The author claims Sage reframes job searching from "applying faster" to "becoming the person those jobs are looking for." This is a repositioning of the job-search problem as one of personal development rather than transactional submission.

Inference This positioning reflects an attempt to differentiate from traditional ATS tools and resume optimizers by focusing on skill gap identification and coaching, not just keyword matching or auto-fill.

The description does not name specific target customers. However, it implies the tool is aimed at job seekers who are:

• Looking for roles in technical fields (AI/ML, Backend, Data Engineering)
• Wanting to improve their competitiveness rather than just submit more applications
• Interested in personalized development paths based on real job requirements

Inference The ICP likely includes recent graduates or early-career professionals seeking to transition into specific domains.

No business model or pricing information is provided. The description does not mention monetization, subscriptions, or any commercial structure beyond the author’s own development effort.

Not evidenced

The project was built with:

• Vite, TanStack Start, TypeScript
• React frontend with Tailwind CSS
• Node.js backend for model calls
• Codex and OpenAI GPT-5.6
• Structured JSON schema prompting
• Client-side resume parsing (pdf.js, mammoth)
• Server boundary to protect API keys

Inference The architecture is designed around verification and structured outputs rather than simple retrieval or similarity matching. It uses type systems and code-level checks to enforce correctness.

There is no evidence of traction, revenue, customers, or adoption beyond the author’s own account. The project was submitted to a hackathon and has no stated deployment, user base, or usage metrics.

Not evidenced

The description does not reference competitors directly. However, it positions itself as an alternative to:

• Traditional job boards (e.g., LinkedIn, Indeed)
• Resume optimizers (e.g., ResumeWorded, Jobscan)
• ATS tools that focus on matching keywords or optimizing applications

Inference Sage aims to differentiate by offering a more nuanced, reasoning-based approach to skill gap analysis and coaching.

1. No evidence of real-world usage or feedback: The tool is described only as a hackathon project with no external validation.
2. Unproven scalability: The dataset used for training/analysis is small (18 postings) and curated, not representative of broader market trends.
3. Model dependency without robust safeguards: Despite using structured outputs, the system still relies heavily on LLM behavior, which can hallucinate or misinterpret data.
4. Limited dataset convergence: While the author claims recurring requirements exist within role types, this is not independently verified.
5. Self-reported accuracy: The verification layer is described as a safeguard but is not shown to have been tested in practice.

1. What real-world testing or feedback has been gathered from users?
2. How does the tool handle edge cases where resume parsing fails or job descriptions are ambiguous?
3. Has the dataset been validated for diversity and representativeness across industries, seniorities, and regions?
4. Are there plans to scale beyond the current hackathon prototype?
5. What is the long-term vision for monetization or product evolution?