Screen Loop

The description states that Screen Loop takes a job description and batch of resumes, then runs three GPT-5.6 agents in sequence:

1. Parser agent: structures each resume.
2. Evaluator agent: scores candidates requirement-by-requirement, citing verbatim quotes from the actual resume for each judgment.
3. Advocate agent: reviews bottom-half rankings to argue for candidates rejected by stricter filters; if the Advocate’s case is strong, the candidate is re-scored.

The system outputs a ranked shortlist where every score traces back to evidence a recruiter can inspect or defend. One click drafts personalized outreach; reasoning trails export as JSON or Markdown.

Inference: The tool is built around structured LLM workflows and emphasizes auditability over traditional AI ranking.

The author claims Screen Loop addresses issues in current AI recruiting tools:

• Black-box scoring
• Lack of defensibility to hiring managers or candidates
• Keyword filter bias that rejects formatting-based candidates

It positions itself as a solution to these problems by introducing:

• Evidence-based scoring with verbatim resume quotes
• An Advocate agent to rescue overlooked strong candidates
• Transparency in decision-making

Inference: The positioning is framed around trust and defensibility, not just automation or speed.

The description states that the author runs agency-side recruiting and Meta Ads work alongside a CS degree. Outreach is already underway with recruiting agencies as part of separate agency work.

Inference: The target customer appears to be recruiting agencies, though no explicit customer list or segmentation data is provided.

Not evidenced. The description does not mention pricing models, monetization strategies, or any commercial structure beyond the author’s own use case and outreach efforts.

• Built with: asyncio, codex, fastapi, gpt-5-6, nextjs, openai-api, pydantic, pypdf, pytest, python, react, server-sent-events, typescript.
• Uses GPT-5.6 for structured requirement extraction and agent workflows.
• Backend never trusts model self-reported scores; final score computed server-side from verified, weighted requirement coverage.
• Engineering decision: no agent framework, just a plain Python orchestrator so every handoff stays readable in one sitting.
• PDF/TXT extraction layer written by Codex.
• SSE streaming for progress updates.
• Synthetic test set generation and pytest suite also scaffolded by Codex.

Inference: The tool is built with a focus on readability, control, and auditability of LLM outputs, not scalability or enterprise-grade infrastructure.

Not evidenced. No revenue data, customer names, usage metrics, or adoption indicators are provided beyond the author’s own outreach efforts and internal testing.

The description does not provide any information about competitors or market positioning relative to existing AI recruiting tools.

1. No traction evidence: The project is described as a hackathon submission with no verified customers or revenue.
2. Single-person team: Only one member listed (Muhammad Burhan), which raises questions about scalability and operational capacity.
3. Self-reported tooling: All technical claims are self-reported without independent verification.
4. Lack of pricing or monetization strategy: No indication of how the product will be monetized or whether it has a sustainable business model.
5. Unverified claims: The author states that “the most valuable audit isn’t 'does the model reason well'; it's 'does literally every input this system might see get handled'”, but no evidence is provided that this has been tested or validated.

1. What specific recruiting agencies are you reaching out to, and what feedback have they given?
2. How do you plan to scale beyond the single-person development team?
3. Are there any early adopters or pilots currently underway?
4. What is your go-to-market strategy for acquiring customers in the recruiting space?
5. How do you intend to monetize this tool? Is there a pricing model in place?
6. Can you walk us through how the Advocate agent makes decisions and what criteria it uses?
7. Have you identified any regulatory or compliance concerns related to handling resumes and candidate data?