GenAI Engineer (Banking/Financial)

Unison Group; Singapore; On-site; SGD 180,000 - 240,000; Full time

Job summary

Unison Group is seeking an experienced software/data engineer to lead LLM-driven solutions in a regulated financial services context. You will design and implement secure AI-enabled pipelines, ensure robust data privacy and access controls, and advance Text-to-SQL with conversational KPI retrieval.

The role emphasizes applying governance, CI/CD, and tooling like LangGraph/LlamaIndex, with a focus on production-ready AI systems in banking environments.

Qualifications

• 8+ years in software or data engineering, including 2+ on LLM applications and 3+ in banking or regulated financial services.
• Experience in building software in regulated environments: authentication, entitlements, audit logging and data privacy.
• Experience with LLM applications beyond proof of concept: RAG, embeddings, vector stores and structured outputs.
• Text-to-SQL over a semantic layer; conversational retrieval of reports and KPIs.
• Agent frameworks such as LangGraph or LlamaIndex.
• Evaluation: accuracy, SQL correctness, hallucination testing and guardrails, automated in CI/CD.
• Code parsing and lineage extraction for SQL, PL/SQL, SAS and Python (e.g. sqlglot, ANTLR).

Responsibilities

• Build software in regulated environments with proper authentication, entitlements, auditing and data privacy controls.
• Apply row-level access controls and human review steps to AI outputs.
• Develop and deploy LLM applications beyond PoC, including RAG, embeddings and vector stores.
• Implement Text-to-SQL over semantic layers and enable conversational KPI/report retrieval.
• Leverage agent frameworks such as LangGraph or LlamaIndex.
• Set up evaluation pipelines for model accuracy, SQL correctness and guardrails in CI/CD.
• Implement code parsing and lineage extraction for SQL/PL-SQL/SAS/Python using tools like sqlglot/ANTLR.

Skills

Python

SQL

LLM applications

Text-to-SQL

Data privacy

Tools

Git

CI/CD

MLflow

sqlglot

ANTLR

LangGraph

LlamaIndex

Job description

Experience: 8+ years in software or data engineering, including 2+ on LLM applications and 3+ in banking or regulated financial services

Must-have skills

Banking and controls

• Building software in regulated environments: authentication, entitlements, audit logging and data privacy.
• Applying row-level access and human review steps to AI outputs.
• LLM applications beyond proof of concept: RAG, embeddings, vector stores and structured outputs.
• Text-to-SQL over a semantic layer; conversational retrieval of reports and KPIs.
• Agent frameworks such as LangGraph or LlamaIndex.
• Evaluation: accuracy, SQL correctness, hallucination testing and guardrails, automated in CI/CD.
• Code parsing and lineage extraction for SQL, PL/SQL, SAS and Python (e.g. sqlglot, ANTLR).

Software and data engineering

• Strong Python and SQL.
• Working with enterprise data at warehouse scale.
• Git, CI/CD and MLflow.

Good-to-have skills

• Databricks Mosaic AI or Genie.
• Visualisation libraries (Plotly, ECharts, Vega) and Qlik or Power BI APIs.

Certifications (preferred): Databricks Generative AI Engineer Associate; Azure AI Engineer; AWS ML Specialty.