About MincaAI

MincaAI is an AI-native insurance software company. HQ at Station F (Paris) x Mexico, operations across LatAm. We build production AI systems for enterprise insurance carriers across three pillars — underwriting, policy issuance, and claims. Real LLM pipelines. Real enterprise clients.

We are small, we move fast, and we ship into environments where being wrong is expensive. The problem

Insurers and brokers exchange data as free text written by humans. No schema, no consistency. Before anything can be priced, quoted or underwritten, that text has to be resolved against large standardized industry catalogs — with a confidence score an underwriter acts on without double-checking it.

This is entity matching and ranking, at scale, in Spanish, in a regulated domain where a confidently wrong answer is expensive. What you will do

Design hybrid retrieval pipelines combining semantic, lexical, fuzzy, and structured search.

Build multi-stage ranking systems using embeddings, rerankers, and LLMs where they add measurable value.

Develop confidence calibration to determine when predictions should be auto-accepted or reviewed by humans.

Build evaluation pipelines, benchmark datasets, and experiment frameworks to measure quality improvements.

Deploy and operate production ML services with a focus on latency, reliability, and cost.

Work closely with insurance experts to continuously improve model performance using production feedback.

What we are looking for Production AI & Machine Learning

5+ years of experience designing, building, and deploying production AI or Machine Learning systems.

Demonstrated ownership of an end-to-end ML system,



from data and model development through deployment, monitoring, and continuous improvement.

Experience owning measurable quality metrics (e.g. search relevance, matching accuracy, recommendation quality) and using data to drive model improvements.

Active Learning or Human-in-the-Loop ML workflows.

Information

Retrieval & Search (Core Requirement)

Strong hands-on experience designing search or retrieval systems using a combination of semantic, lexical, and structured retrieval techniques.

Deep understanding of candidate generation, retrieval pipelines, and ranking architectures.

Experience evaluating and improving retrieval performance using metrics such as Recall@K, Precision@K, MRR, or nDCG.

Ability to reason about retrieval trade-offs, including recall vs. precision, latency vs. quality, and dense vs. lexical search.

Vector databases such as pgvector, Pinecone, Milvus, Weaviate, or Qdrant. Ranking & Entity Matching

Experience building ranking, entity resolution, record linkage, recommendation, or similar matching systems.

Familiarity with hybrid ranking approaches combining embeddings, lexical signals, fuzzy matching, and business rules.

Experience with multi-stage retrieval and reranking pipelines is highly desirable.

Learning-to-Rank or Cross-Encoder reranking.

Large Language

Models





Production experience integrating LLMs into AI workflows beyond prompt engineering.

Experience with structured outputs, function/tool calling, prompt evaluation, model selection, caching, and cost optimisation.

Strong understanding of where LLMs add value—and where traditional ML or search techniques are more appropriate.

Experience fine-tuning embedding models or domain adaptation. Evaluation & Experimentation

Experience designing evaluation datasets and benchmarking frameworks for ML systems.

Strong analytical skills in error analysis, model comparison, regression detection, and experiment design.

Familiarity with offline evaluation and validating improvements before production deployment.

Software

Engineering

Strong Python development experience, including writing maintainable, well-tested, production-quality code.

Experience building APIs using FastAPI or similar frameworks.

Experience working with PostgreSQL, including query optimisation and indexing.

Familiarity with Docker, Kubernetes, Git, and modern CI/CD workflows. Nice to Have

Confidence calibration techniques (e.g.

Platt

Scaling, Isotonic Regression).

Experience in insurance, financial services, healthcare, or other regulated industries.

Spanish language proficiency (preferred), with French or Vietnamese considered a plus. Logistics

Location: Remote, Latam Languages: Professional English + Spanish required. French a plus.

Timezone: Overlap with LatAm working hours Start: ASAP. How to Apply

Send your CV + a short note (5 lines max) to [email protected] answering : "what do you think you are the best candidate"

📌 # AI Engineer - Search, Matching & Ranking (Argentina)
🏢 MincaAI
📍 Argentina