AI/ML Architect - Banking
Unison GroupSingapore, Singapore, SingaporeOn-Site
Needs a free Superjobs account. You'll finish on Unison Group's site.
Skill score
Summary
The AI/ML Architect designs and governs production AI and GenAI solutions inside banks, covering RAG, text-to-SQL over semantic layers, LLM evaluation, and code intelligence over legacy SQL, PL/SQL, SAS and Python. The role suits a senior practitioner with a decade of data work, several years of ML delivery in banking, and experience leading multidisciplinary teams and presenting controls to model risk and audit.
- Function
- Data Science
- Estimated salary
- Est. $101k–151k/yr
- Estimated for a staff-level role, ~10 yrs experience, Singapore. Not published by the employer.
- Work mode and location
- On-Site, Singapore, Singapore, Singapore
- Experience
- 10+ yrs
- Employment type
- Contract
Job description
Overview
We are looking for AI/ML Architect with 10+ years in data and analytics, including 6+ in ML/AI, 3+ in banking and 2+ on production GenAI
Skills required
Top skills
- ML/AI
- Generative AI
- RAG
- Text-to-SQL
All skills (12)
- Model risk governance
- Responsible AI
- Financial crime data
- Python
- Spark
- TensorFlow
- PyTorch
- MLflow
- Hadoop
- LangGraph
- Vector stores
- Databricks
Qualifications
Banking and AI governance
- Taking AI/ML solutions to production inside banks through model risk or AI governance review.
- Responsible AI: explainability, audit trails, human-in-the-loop controls and model monitoring (e.g. MAS FEAT principles).
- Financial crime data: AML models, alerts and case data.
Generative AI architecture
- RAG, text-to-SQL over semantic layers, and agentic analytics for insight generation, metadata discovery and query generation.
- Integration with enterprise LLM platforms, enforcing entitlements, row-level security and prompt/response audit.
- LLM evaluation: accuracy, SQL correctness, hallucination testing and guardrails.
- Code intelligence: parsing SQL, PL/SQL, SAS and Python to extract mappings, lineage, dependencies and business rules.
- Generating specifications, code and tests from metadata, with traceability.
ML Engineering
- Python, Spark, Scala or Java; TensorFlow, PyTorch and scikit-learn.
- MLflow, feature engineering, model deployment and monitoring; DataOps practices.
- Large-scale data on Hadoop and Lakehouse platforms; semantic layers, knowledge graphs and federated governance.
Delivery and communication
- Leading multidisciplinary teams across data engineering, analytics, AI/ML and business domains.
- resenting AI designs and controls to model risk, audit and business stakeholders.
Nice to have
- Agent frameworks such as LangGraph; vector stores.
- Enterprise LLM platforms: Azure OpenAI, AWS Bedrock, Vertex AI or on-premises models.
- Databricks Mosaic AI, Genie and Unity Catalog; SQL parsers such as sqlglot.
Certifications (preferred): Databricks Generative AI Engineer or ML Professional; Azure AI Engineer; AWS ML Specialty; Google Professional ML Engineer.
Company info
About Unison Group
Unison Group provides digital transformation services to enterprises, offering technology solutions across fintech, AI, risk management, analytics and cloud. Its clients are large organizations in finance, technology and consulting, including regulated multinational businesses.