DESCRIPTION
The AI Automation team is in charge of generating all the AI related tools that the company needs while scaling. The team is small (3 people) and every member owns a subset of services, end-to-end. The team is currently managing chat-bots, voice-bots, document analysis, customer verifications and communications monitoring.
Aviva's loan origination relies on a real-time evidence verification system that processes identity documents, geolocation, videos, and bank evidence to make compliance decisions in seconds. You will own the engineering and continuous improvement of this system — from ML-based document analysis to the business rules that turn extracted signals into a disbursement or internal approvals.
This is a hands-on role. You will write production code, design new verification flows, and
collaborate directly with the compliance, risk, and product teams to translate business needs and regulatory requirements into software.
The customer verification platform is a multi-service Python system that sits on the critical path of every loan application. It runs 11 concurrent check types (identity, address, bank account, geolocation, videos, contract, interview, and more) against customer-submitted evidence.
Evidence flows through typed analyzers (image, video, contract, JSON) into specialized
sub-analyzers — INE card OCR, address extraction, bank statement parsing, signature verification. Sub-analyses produce structured traits (with ML probability scores) that feed into check business rules, which emit structured appraisal items and human-friendly feedback messages.
-
Design and implement new checks end-to-end: evidence definition evidence-analysis verification checks structured feedback tests.
-
Add new fraud signals: document tamper detection, liveness scoring, synthetic identity patterns, voice matching, and face matching.
-
Products tend to reuse componentes but often they need new evidences, analysis, verifications, etc.
-
Integrate external verification services (e.g. ID registries, identity matching services) as async clients with proper timeout and error isolation.
-
Implement automated rejection flows so compliance decisions that meet defined confidence thresholds require no human review.
-
Manage the feedback pipeline: produce machine-readable appraisal items for internal consumers and dynamic human-friendly messages for applicants.
-
Design a structured exception hierarchy that lets downstream systems handle errors precisely (retriable vs. permanent, evidence-level vs. check-level).
-
Drive incremental refactors of the analysis engine to reduce coupling between analyzers, sub-analyzers, and checks — keeping each layer independently deployable and testable.
-
Maintain high unit-test coverage for business rules and integration tests for the full evidence-to-decision flow.
-
Monitor check pass rates, appraisal rejection reasons, and external service latency via Dagster pipelines and BigQuery dashboards.
-
Own the on-call escalation path for verification failures that block disbursements.
-
The main consumer of these architecture is the Aviva Operating System, so extensive tests and coordination are required when there are changes in schemas or endpoints.
- Versioning endpoints with new features to decouple the deployments .
-
Attractive compensation package.
-
Fast-paced environment with significant growth opportunities.
-
15 annual vacation days + 7 annual personal days.
-
Option to work remotely 3-4 days per week ; or fully-remote (as long as you can come to CDMX ~twice a year)
-
Flexible work schedule
REQUIREMENTS
-
5+ years of Python development in production systems.
-
Containerisation & orchestration — Hands-on experience with Docker and Kubernetes in a production setting (resource management, rolling deployments, health probes).
-
Strong async Python: you understand event loops, structured concurrency, and how to reason about distributed work queues (Celery, Redis, or equivalent).
-
Ability to design extensible, type-safe systems using Pydantic, dataclasses, and generics.
-
Experience owning production services that are consumed by multiple teams.
-
Ability to read business and regulatory requirements and translate them directly into code.
-
Monitoring & observability — Experience designing dashboards, alerts, and distributed traces for services where "the service returned 200 but the answer was wrong" is a real failure mode.
-
Communication — You will be the main point of contact between Data Science and the engineering teams for a subset of services. Clear, precise written and verbal communication is essential.
-
Fluency in both Spanish and English. Most of our meetings are in Spanish, but the code and most documentation is written in English.
-
Experience with document intelligence APIs (Google Cloud Document AI, Azure Form Recognizer or similar OCR pipelines).
-
Experience with computer vision, fine-tunning image classification models, labeling, testing, full pipelines. Use of OCR alchemy
-
Experience with LLMs at production level. Use of AILabs for testing chatbots
-
Knowledge of the Azure ecosystem (AKS, ACR, Azure DevOps).
-
Familiarity with API-testing tools such as Bruno or Postman for contract and integration testing.
-
Familiarity with Pants, or other similar build systems.
