Company Description
We are Software Mind, an awesome team of engineers who are ready to ramp up any top-notch company’s projects! Our aim? To always be one step ahead. Become part of a multicultural company in constant growth with an excellent work environment certified by Great Place To Work!
Job Description
About the Project
Software Mind is building a private, tenant-isolated AI assistant for the real estate title and settlement industry. The platform is a retrieval-first (RAG) system that ingests historical email, documents, and structured metadata into a per-tenant vector index, and serves grounded, cited, expert-weighted answers through a chat-style Q&A interface with single sign-on and full audit logging.
The platform is AWS-native with a Python/FastAPI backend, Vue.js frontend, OpenSearch/Pinecone vector store, and OpenAI/Anthropic/Bedrock as LLM provider. You will join a senior, cross-functional LATAM-based team where hands-on AI delivery experience not just familiarity is the baseline expectation.
You will serve as the requirements and alignment anchor for the engagement translating domain knowledge, SME expertise patterns, and retrieval accuracy logic into structured, traceable specifications the engineering team can build against. You will also design the accuracy validation rubrics used at go/no-go gates. This is not a documentation-only role.
Your Responsibilities
Lead discovery sessions and SME alignment workshops; produce structured requirements and integration assumption documentation
Define data ingestion scope: folder selection, corpus boundaries, PII classification requirements, and tenant governance rules
Translate expert-weighting business logic into specifications for the retrieval and ranking layer
Author and maintain the accuracy rubric framework used for SME validation
Write and validate user stories (INVEST-compliant) with acceptance criteria tied to delivery milestones
Produce and maintain schema documentation for the data processing pipeline
Facilitate query clustering analysis to prioritise accuracy tuning work
Bridge between client Designated Expert input and the engineering team's implementation backlog
Maintain requirements traceability across Jira and Confluence throughout all phases
Tech Stack: Jira, Confluence, Microsoft 365, Miro, FigJam.
Qualifications
Must-Have Skills & Experience
+90% English written and oral (at least B2 level). Excellent written and spoken English direct engagement with client stakeholders is expected
6+ years as a Business Analyst on software delivery projects; senior seniority required
Proven ability to translate complex business needs into structured, developer-ready requirements (user stories, use cases, acceptance criteria)
Strong data literacy: ability to reason about document structures, metadata schemas, and classification logic
Experience eliciting requirements from domain SMEs and translating them into implementable specifications
Proficiency with user story authoring, backlog management, and acceptance criteria definition in agile environments
AI Experience (Required Not Optional)
At least one project involving requirements definition or evaluation design for a production AI/ML or NLP system RAG, search, classification, recommendation, or similar
Ability to define evaluation criteria and accuracy rubrics for AI system outputs not just functional requirements
Understanding of LLM behaviour, grounding, hallucination risk, and retrieval relevance trade-offs at a conceptual level sufficient to write testable acceptance criteria
Experience bridging between domain subject-matter experts and AI/ML engineering teams
Additional Information
Nice-to-Have
Domain exposure to legal, compliance, financial services, or document-heavy regulated industries
Experience with PII classification, data governance, or privacy-by-design requirements
Familiarity with LangChain, vector search concepts, or document processing pipelines at a conceptual level
We are accepting applications from LATAM countries