INTL
Freelancer
전문가
외주
원격 가능
Vacancy search & MVP AI Agent
예산
$750~$1,500 EUR
예상 기간
1~2개월
난이도
전문가
기술 스택
Python
LangChain
OpenAI API
Web Scraping
Vector Database
FastAPI
Docker
RAG
AI 분석 요약 기회점수 55/100
본 프로젝트는 후보자의 CV, LinkedIn 프로필, Metaview 인터뷰 기록을 통합 분석하여 최적의 채용 공고를 찾아 매칭하고, 채용 담당자에게 초기 이메일을 자동 생성 및 발송(또는 스크립트 제공)하는 MVP AI 에이전트를 개발하는 과제입니다. 복수의 데이터 소스 통합, 고급 LLM 기반 정보 추출 및 웹 스크래핑, 그리고 GDPR 및 개인 정보 보호 규정 준수가 핵심 요구사항입니다. AI/ML, NLP, 웹 스크래핑 및 LLM 오케스트레이션(LangChain) 경험이 풍부한 숙련된 개발자에게 적합합니다.
핵심 요구사항
- 후보자 CV, LinkedIn, Metaview 데이터를 통합 및 정규화하여 관리
- LLM/NLP 기반으로 스킬, 경력, 선호도 추출 및 최적의 채용 공고 매칭
- 기업 채용 페이지 및 채용 게시판에서 관련 채용 공고 검색 및 순위 매기기
- GDPR 및 개인 정보 보호 규정, LinkedIn/잡보드 약관 준수
- MVP 이메일 자동 생성 및 발송(또는 스크립트) 기능 구현
- Python, LangChain, OpenAI API 등 최신 AI/ML 스택 활용
산출물
- 로컬 실행 가능한 스크립트 또는 마이크로서비스 (CLI/간단한 웹 UI)
- 최소 3개 가상 후보자에 대한 샘플 JSON/CSV 출력
- 설정, 환경 변수, 규칙 재학습/미세 조정 방법을 설명하는 README 문서
매력 포인트
- LangChain, OpenAI, LLM, RAG 등 최신 AI/ML 기술 스택 활용 경험 기회
- 명확한 요구사항, 산출물 및 성공 기준 명시
- 다양한 데이터 통합 및 지능형 매칭 시스템 개발의 흥미로운 문제 해결
주의사항
- 요구되는 전문성과 복잡성 대비 매우 낮은 예산 (1~2개월 $750~$1,500 EUR)
- 광범위한 규정 준수 요구사항(GDPR, 개인 정보 보호, LinkedIn/잡보드 약관)에 따른 높은 법적/기술적 위험
- Metaview 통합 난이도 및 API 접근성에 대한 불확실성
프로젝트 유형신규개발
적합 개발자5년차 이상 AI/ML 엔지니어 또는 LLM 아키텍트 (Python, LangChain, Web Scraping)
프로젝트 원문 설명
I’m building an automated Agent that can read a candidate’s CV, LinkedIn profile, and the transcripts or summaries produced by Metaview, then instantly surface the most relevant open roles for them from across company career pages and job boards.
Here is what I need the agent to handle end-to-end:
• Ingest the three data sources (CV, LinkedIn profile, Metaview interview notes) and keep them linked to the same candidate record.
• Extract and normalise skills & competencies, work experience, plus any stated or implied preferences for the ideal next role. The skills and experience must be pulled from every data source, not just one.
• Use the consolidated profile to query company career pages, job boards or an internal vacancy API, rank matches by fit, and output a concise shortlist with an explanation of why each role is a match.
• Exclude vacancies from recruitment agencies, executive search and headhunting companies
• identify vacancy contact / hiring manager where possible
• Automate production of MVP email to the vacancy contact / hiring manager, with anonymised details of candidates. Where unable to automate the send, automate the script for me to send manually via my LinkedIn
• Compliant with GDPR and Privacy regulations across the US, UK and Europe
• Compliant with LinkedIn and job board terms and conditions
• Cost effective in terms of data processing, storage and CPU, balanced against performance
Technology choices are up to you, but I expect modern NLP / LLM tooling (Python, LangChain, OpenAI, Claude Code or similar) and clear, well-commented code so the pipeline can be expanded later.
Deliverables
– Working script or micro-service that runs locally (CLI or simple web UI is fine).
– A sample JSON or CSV output for at least three dummy candidates.
– Brief README explaining setup, environment variables (API keys), and how to retrain or fine-tune extraction rules if needed.
Acceptance criteria
1. All three data sources are successfully parsed and merged.
2. Skills, work experience and preferences are correctly extracted and appear in the output.
3. Vacancy shortlist clearly orders roles by relevance with a short rationale.
4. Setup from clean machine to first run takes under 15 minutes following the README.
If you already have experience with résumé parsing, web scraping, LinkedIn or Metaview integrations, that will jump you to the front of the line. I’m ready to start as soon as I find the right partner.
No vibe coders without prior software development and architecture experience please.
Here is what I need the agent to handle end-to-end:
• Ingest the three data sources (CV, LinkedIn profile, Metaview interview notes) and keep them linked to the same candidate record.
• Extract and normalise skills & competencies, work experience, plus any stated or implied preferences for the ideal next role. The skills and experience must be pulled from every data source, not just one.
• Use the consolidated profile to query company career pages, job boards or an internal vacancy API, rank matches by fit, and output a concise shortlist with an explanation of why each role is a match.
• Exclude vacancies from recruitment agencies, executive search and headhunting companies
• identify vacancy contact / hiring manager where possible
• Automate production of MVP email to the vacancy contact / hiring manager, with anonymised details of candidates. Where unable to automate the send, automate the script for me to send manually via my LinkedIn
• Compliant with GDPR and Privacy regulations across the US, UK and Europe
• Compliant with LinkedIn and job board terms and conditions
• Cost effective in terms of data processing, storage and CPU, balanced against performance
Technology choices are up to you, but I expect modern NLP / LLM tooling (Python, LangChain, OpenAI, Claude Code or similar) and clear, well-commented code so the pipeline can be expanded later.
Deliverables
– Working script or micro-service that runs locally (CLI or simple web UI is fine).
– A sample JSON or CSV output for at least three dummy candidates.
– Brief README explaining setup, environment variables (API keys), and how to retrain or fine-tune extraction rules if needed.
Acceptance criteria
1. All three data sources are successfully parsed and merged.
2. Skills, work experience and preferences are correctly extracted and appear in the output.
3. Vacancy shortlist clearly orders roles by relevance with a short rationale.
4. Setup from clean machine to first run takes under 15 minutes following the README.
If you already have experience with résumé parsing, web scraping, LinkedIn or Metaview integrations, that will jump you to the front of the line. I’m ready to start as soon as I find the right partner.
No vibe coders without prior software development and architecture experience please.
Freelancer에서 원본 확인
원본 보기