Data engineers for Philippine banks and insurers: what BSP changed
- Recruitment
- 8 min read
BSP’s June 2026 AI governance principles are voluntary, but they keep boards accountable for AI and name data quality, bias, privacy and data poisoning as risks to control. For Philippine banks and insurers, that turns data engineers, data quality leads and model owners into governance roles, and hiring should test judgement, not just tools.
What did BSP issue on AI in 2026, and is it binding?
Not yet. On 30 June 2026, according to press reports, the Bangko Sentral ng Pilipinas (BSP) released Memorandum No. M-2026-031, Governance Principles for Artificial Intelligence in Financial Services, and described it as voluntary and non binding. BSP also said further binding rules may follow, which is the sentence every CRO should underline.
The principles are built on STARS: Sustainability, Transparency, Accountability, Responsibility and Security. Boards and senior management remain accountable for AI. Controls should be proportionate to each system’s complexity and materiality. The memorandum covers bias, data quality, privacy and data poisoning.
Days later, on 6 July 2026, BSP Memorandum No. M-2026-034 addressed AI enabled cyber threats: asset inventories, stronger authentication, lower internet exposure, AI based threat detection, incident response and business continuity. Then on 27 July 2026 the BSP AI Summit ran under the theme “From Pilot to Practice: AI Adoption in the Philippine Financial Sector”. Read together, the direction is clear. BSP wants AI in production, with named owners and clean data underneath it.
Which roles do BSP’s AI principles imply?
They imply owners. A voluntary principle that boards remain accountable only works if someone below the board can show, system by system, what the AI does, what data it uses and how it is controlled. If your version of these people is spread across IT, risk and analytics with nobody connecting them, that is the gap to close first.
Theme in the BSP memoranda | Role that usually owns it |
Board and senior management accountability | AI governance lead reporting into the CRO |
Proportionate controls by complexity and materiality | Model risk manager who tiers AI systems |
Data quality | Data quality lead, supported by data engineers |
Bias | Model owners and validators who test outcomes across customer groups |
Privacy | Privacy lead working alongside data engineering |
Data poisoning and AI enabled cyber threats | Security engineers who understand machine learning |
Open finance adds pressure on the same teams. BSP’s open finance framework dates from Circular No. 1122 in 2021, with a pilot running since 2023, and sharing data safely with third parties is pipeline, API and access control work. That is squarely a cloud and data engineering recruitment problem, and it competes for the same engineers your AI programme needs.
What should Philippine banks and insurers test in data interviews?
Test whether the candidate can make a number trustworthy, not whether they can recite a tool list. SQL on messy data is the floor. Above it, look for reconciliation, lineage and the instinct to ask where a figure came from before building on it.
A practical interview has three parts. First, a SQL exercise on two extracts that do not agree, where the candidate has to find and explain the break. Second, a data quality judgement: here is a customer table with duplicates and missing fields, what do you fix, what do you flag, and who do you tell? Third, for AI roles, ask them to explain one model decision to a branch manager in plain language. The STARS principle of transparency is a hiring criterion as much as a policy.
The discipline behind ETL and data engineering in insurance and banking is the same in Manila as anywhere else. Candidates who have closed books, reconciled a ledger or cleaned a claims file under deadline are usually worth more than a longer list of certificates.
How do analysts move into data engineering in the Philippines?
The usual route is from analyst or reporting roles inside the bank or insurer, and it works because those people already know the data. IBPAP, the industry association, reported 2025 industry revenue of US$40.3 billion and about 1.89 million full time equivalent workers. That is context, not a count of banking data talent, but it points to a large workforce already used to handling operational data at volume. The step up is engineering habit.
For the analyst, that means owning one pipeline end to end, writing tested code instead of spreadsheet macros, learning version control and scheduling, and taking on the data quality work others avoid. For the employer, it means creating the path deliberately. Pair analysts with engineers on real month end loads, give them a reconciliation job to own, and assess them on it.
Treat analytics and AI recruitment and internal development as one plan. The analyst you promote and the engineer you hire should be scoped together, or you end up with two people doing half of the same job.
Can data roles be hybrid or remote for foreign employers?
Often yes, but design the access model before you promise the working model. M-2026-034 asks for asset inventories, stronger authentication and lower internet exposure. A remote data engineer with production access to customer data is exactly the exposure a security team will want inventoried, authenticated and logged.
For a regional or foreign firm with a Manila team, the workable pattern is hybrid with controlled environments: data stays in governed platforms, engineers work through secured virtual desktops or managed devices, and production changes go through the same approvals as for office staff. Take local legal advice on the employment structure and on data transfers before you advertise a fully remote role.
If you are still deciding where to place the team at all, that is a separate question, and India or the Philippines for an insurance capability centre is a decision to make on the work itself, not on headline costs.
Should AI screen data engineering candidates?
AI must not replace human judgement in screening. Candidates now write their CVs with the same tools recruiters use to read them, so automated matching errs both ways: it rejects the reconciliation specialist whose CV undersells her and passes the polished profile with nothing behind it. Where AI earns its place in hiring is on a firm’s own recruitment data, meaning interview notes, feedback and hiring manager preferences, not on an isolated CV. If a bank would not accept an unexplained decision in lending, it should not accept one in hiring either.
Frequently Asked Questions
Is BSP's AI governance memorandum binding on banks?
No. BSP Memorandum No. M-2026-031, Governance Principles for Artificial Intelligence in Financial Services, released on 30 June 2026, is voluntary and non binding. BSP said further binding rules may follow, so banks, insurers and other supervised institutions should treat the principles as the likely shape of future requirements and start assigning owners and controls now.
What does STARS stand for in BSP's AI principles?
STARS stands for Sustainability, Transparency, Accountability, Responsibility and Security. These five principles underpin BSP Memorandum No. M-2026-031, released on 30 June 2026. The memorandum keeps boards and senior management accountable for AI, asks for controls proportionate to each system’s complexity and materiality, and covers bias, data quality, privacy and data poisoning.
Will AI replace entry level data analyst jobs in Philippine banks?
AI will absorb much of the routine report building that entry level analysts do today. It will not remove the need for people who check data, question outputs and explain results, which is what BSP’s emphasis on data quality and accountability requires. Analysts who add engineering and data quality skills move toward the work that is growing.
What did BSP Memorandum M-2026-034 cover?
Issued on 6 July 2026, BSP Memorandum No. M-2026-034 addressed AI enabled cyber threats. It covered asset inventories, stronger authentication, lower internet exposure, AI based threat detection, incident response and business continuity. For data teams, it affects how remote access, production credentials and data platforms are designed and documented.
If you are building a data or AI team for a Philippine bank or insurer and want a specialist view on the brief before you go to market, Start a conversation.
Talk to us about your hiring
Get in touch →More reading
Related intelligence
-
Recruitment
ETL developers in insurance and banking: what they do, how to hire
An ETL developer or data engineer in an insurer or bank moves policy, claims, transaction and ledger data…
Read the article → -
Recruitment
How internal audit should audit AI: skills, standards and team design
Internal audit should audit AI the way it audits any high risk process: test governance, data, validation, monitoring…
Read the article → -
Recruitment
Insurance data, AI and actuarial hiring in Japan after ESR
Japan's economic value based solvency regime and the FSA's work on AI have raised demand for actuaries, modellers…
Read the article →