Reljod OretaContact

AI Consultant · Banking and financial services

AI for banks, built to ship

I help banks and financial firms turn slow, manual compliance work into AI services that run in production: KYC, document checks and name matching. Seven years in software, the last two shipping multi-agent LLM systems for a digital bank and multi-tenant SaaS platforms.

Book a callSee the work

What I do

Compliance work, automated

Banks run on checks: identity, documents, names. I build the AI systems that take the repetitive part off your team and hold up to review.

  • KYC automation

    In-house AI identity verification that replaces paid third-party vendors and lifts approval and automation rates.

    • Identity document extraction
    • Approval and review routing
    • Vendor replacement plans

  • Document compliance checks

    AI pipelines that review submitted documents against your compliance requirements before a person has to.

    • Requirement checklists as code
    • Tampering detection
    • Illegibility detection

  • Name matching

    Matching names across documents and records to support identity verification and screening.

    • Cross-document matching
    • Fuzzy and transliterated names
    • Explainable match scores

  • AI agents and workflows

    Multi-agent systems built into your existing applications and data platforms, not a parallel stack.

    • Python FastAPI services
    • AWS Lambda, EKS and Bedrock
    • Shared agent framework

  • AI adoption and governance

    A practical start: a few visible early wins, tools like Claude and ChatGPT in people's hands, and a small team governing their use.

    • Current and desired state mapping
    • Early-win workflow picks
    • Usage policy and governance

  • Evaluation and path to production

    The evaluation harness, tracing and CI/CD that prove an AI system works and keep it working after launch.

    • LLM evals and guardrails
    • LangFuse and LangSmith tracing
    • Hosting for business-built apps

How I work

From unclear need to shipped system

The hardest part is rarely the model. It's getting the business to say where it is today and where it wants to be. This is how I get there.

  1. 01

    Frame the problem

    Sit with the business, including auditors and operations staff, to map the current and desired state and agree on success criteria first.

  2. 02

    Win early

    Automate one or two manual workflows so the business sees a visible result within weeks, not quarters.

  3. 03

    Build with evidence

    Agentic engineering backed by tests, evaluation checks, tracing and CI/CD, iterated with users in short loops.

  4. 04

    Hand over a path to production

    Integrate with the systems you already run, document it, and leave your team able to own and extend it.

Experience

Seven years, from finance data to digital banking

Three years inside a leading Philippine digital bank, two at a global financial data company, and the last two shipping production AI.

  1. Jul 2026 – Present

    Remote · Australia

    Consulting

    AI Consultant · Acquire Intelligence

    • Senior AI engineer on client engagements, owning each solution from problem framing through architecture, implementation and deployment.
    • Build AI agents and AI-powered services in Python and FastAPI, integrated with clients' existing applications and data platforms.
    • Python
    • FastAPI
    • AI agents

  2. Dec 2025 – Jun 2026

    Remote

    AI SaaS

    Full Stack Python Engineer / AI Engineer · Cloud Employee

    • Designed a multi-agent eligibility pipeline on AWS Lambda that screens out 95 to 97 percent of ineligible records before report generation.
    • Ran asynchronous agent workflows on Celery, Docker and AWS EKS, serving compliance, finance and KYC use cases from one shared agent framework.
    • Turned operational problems from product and business stakeholders into agent designs, with success criteria and evaluation checks defined before each release.
    • CrewAI
    • LangGraph
    • AWS Bedrock
    • LangFuse

  3. May 2025 – Nov 2025

    Remote

    AI SaaS

    AI Engineer · Spark Intelligence, Inc.

    • Built a full-stack multi-tenant AI SaaS product with agent logic in LangChain and LangGraph over GPT, Gemini and Grok.
    • Engineered a fault-tolerant AI pipeline on Modal and Trigger.dev so long-running inference jobs recovered without dropping tenant workloads.
    • Improved output quality with LangFuse tracing, DSPy prompt optimization and DeepEval custom metrics.
    • Next.js
    • Supabase
    • DSPy
    • DeepEval

  4. Mar 2022 – Apr 2025

    Taguig, Philippines

    Banking

    Senior Software Engineer · Maya Philippines, Inc.

    • Partnered with data science to replace a costly third-party KYC vendor with an in-house AI solution: 90% lower KYC operating cost, +7 pts approval rate, +8 pts automation rate.
    • Core engineer on the User Profiles team of a leading Philippine digital bank and wallet, across Node.js, FastAPI, Spring Boot and AWS serverless.
    • Built fault-tolerant data pipelines on Firehose, S3, Aurora and Databricks to feed analytics and model training.
    • KYC
    • AWS
    • FastAPI
    • Databricks

  5. Feb 2020 – Mar 2022

    Taguig, Philippines

    Financial data

    Software Engineer III · FactSet Philippines, Inc.

    • Led the integration of FactSet's Revere finance data, running stakeholder meetings and reporting delivery timelines to management.
    • Migrated legacy Perl and C++ data pipelines to C# ASP.NET, and automated manual data ingestion with an internal tool.
    • C#
    • MSSQL
    • Python
    • Vue.js

  6. Jun 2019 – Feb 2020

    Manila, Philippines

    Hardware

    Prototype Engineer · Yuro Product Development and Design Solutions

    • Built a computer vision prototype for tennis officiating with OpenCV, YOLO and PyTorch.
    • Designed the AWS IoT Core architecture for a connected hardware product.
    • OpenCV
    • PyTorch
    • AWS IoT
BS Electronics Engineering
Polytechnic University of the Philippines, Manila · DOST Scholar, 2014–2019
Licensed Electronics Engineer (ECE)
Philippine Professional Regulation Commission, 2019

Projects

Selected AI systems

Production systems I've designed and shipped. Full case studies are on the way.

  • Digital bank

    Case study soon

    In-house KYC automation

    AI identity verification that replaced a paid third-party KYC vendor for a Philippine digital bank and wallet.

    90% lower KYC operating cost

    • KYC
    • Document AI
    • AWS

    Case study in progress

  • [Client sector]

    Case study soon

    Document compliance checks

    AI pipelines that review submitted documents against compliance requirements before human review.

    [Outcome metric]

    • Compliance
    • LLM agents

    Case study in progress

  • [Client sector]

    Case study soon

    Name matching engine

    Matching names across documents and records to support identity verification.

    [Outcome metric]

    • Identity
    • Matching

    Case study in progress

  • [Client sector]

    Case study soon

    Tampering and illegibility detection

    Flagging altered or unreadable documents before they reach a reviewer's queue.

    [Outcome metric]

    • Computer vision
    • Document AI

    Case study in progress

  • AI SaaS

    Case study soon

    Multi-agent eligibility pipeline

    Agents on AWS Lambda that screen out ineligible records before report generation, cutting wasted inference spend.

    95–97% of ineligible records screened out

    • Multi-agent
    • AWS Lambda

    Case study in progress

  • [Client sector]

    Case study soon

    Marketing intelligence

    AI features that turn marketing-related data into insights a team can act on.

    [Outcome metric]

    • Analytics
    • LLM

    Case study in progress

Contact

Let's map your first AI win

Tell me where your team spends the most manual hours on checks and reviews. I'll come back with where AI fits, what it takes, and how we'd prove it works.

Email meLinkedInGitHub