Jakarta, Indonesia · Open to consulting

I solve hard
business problems
with data & ML.

Lead Data Scientist with 6+ years building production ML systems across fintech, investment, and marketplace. From fraud prevention to growth engines — I own the full loop.

Work together → GitHub ↗ LinkedIn ↗
6+
Years in production ML
1M+
Investors acquired via ML
−20%
Fraud at <200ms latency
+120%
Loan conversion uplift
01

What I solve

⚖️
Risk & Credit Scoring

Dynamic scoring systems embedded into decisioning flows — for leveraged trading, loan origination, and portfolio-level risk appetite.

📈
Growth & Acquisition ML

Propensity models that turn broad prospect pools into high-value cohorts — driving AUM, ticket size, and net new customer volume.

🔍
Fraud & Anomaly Detection

Real-time transaction monitoring — Kafka event streaming, GPU-accelerated inference, sub-200ms latency at production scale.

🤖
LLM & Applied AI

Practical LLM pipelines for support automation, structured decision handling, and intelligent workflows at scale.

02

Experience

Bareksa
2025 – Present
Lead Data Scientist
  • Own data science across growth & risk — technical direction, team capability, end-to-end ML delivery
  • Designing a dynamic credit risk scoring system for leveraged trading, embedded into core decisioning flows
  • Sustained 2× AUM performance against the national baseline amid a 49% quota reduction — reorienting campaign strategy from mass outreach to a propensity-driven model that protected growth where others contracted
  • Leading cross-functional LLM-powered support automation pipeline with backend engineering
Risk ScoringPropensity ModellingLLMExperimentationGCPAirflowMetabase
Bank Mega
2023 – 2025
Customer Banking Specialist — Asst. Manager
  • Doubled loan conversion: 2.6% → 5.9% (+120%) via predictive decisioning in the origination funnel
  • Unlocked ~60K net new customers — improved acquisition conversion rate +20% by distilling 300K external prospects to a 90K high-value cohort, operationalised as the company's primary targeting engine
  • Reduced operational costs 15% through a rigorous experimentation framework
  • Improved revenue scoring accuracy +20% via joint initiative with Visa Analytics — predicting per-customer revenue within a 1-year horizon
Credit IntelligenceLead ScoringA/B TestingPower BITableauAzure
Grab
2020 – 2023
Data Scientist — Regional (Singapore HQ)
  • Built propensity intelligence for OVO Invest — contributed to 1M+ investor acquisitions in year one
  • Grew AUM per user +35% with a personalised investment recommendation engine
  • Productionised real-time fraud detection: Kafka + GPU anomaly detection at <200ms, delivering 20% fraud reduction
  • Operated across Indonesia within Singapore-led regional DS function
Fraud DetectionRecommendationKafkaPyTorchAWSFirebaseAirflow
Grab
2019
Data Science Intern
  • Supported Anti-Money Laundering team — flagging watchlist individuals using NLP techniques
NLPAML
Universitas Gunadarma
2018 – 2019
Lead Laboratory Assistant
  • Led team of lab assistants coordinating AI projects and ensuring timely delivery across the faculty
AI ProjectsTeam Lead
03

Impact by the numbers

Percentage uplift delivered across growth and acquisition initiatives.

04

Tools & stack

ML & Languages
Python SQL PyTorch TensorFlow GPU Inference
Data & Streaming
Apache Kafka Apache Spark Airflow Firebase
Cloud & Infra
GCP AWS Azure On-premise
Analytics & BI
Metabase Tableau Power BI
05

Selected results

2×
AUM performance vs national baseline at Bareksa — national quota was cut 49%, yet the propensity-based targeting model held growth at 2× what the business would have achieved without it.
Bareksa · 2025–Present
+120%
Loan conversion doubled from 2.6% to 5.9% by embedding a predictive decisioning model into the origination funnel.
Bank Mega · 2023–2025
1M+
Investors acquired in OVO Invest's first year, powered by a propensity intelligence layer built for the product team.
Grab · 2020–2023
+35%
AUM per user growth driven by a personalised investment recommendation engine deployed across Indonesia.
Grab · 2020–2023
−20%
Payment fraud reduced via real-time transaction monitoring — Kafka + GPU at sub-200ms latency, production scale.
Grab · 2020–2023
+20%
Acquisition conversion rate improvement — distilling 300K external prospects to a targeted 90K high-value cohort, unlocking ~60K net new customers and adopted as the company's primary targeting engine.
Bank Mega · 2023–2025
+20%
Revenue scoring accuracy improvement — a model predicting per-customer revenue within a 1-year horizon, built in partnership with Visa Analytics.
Bank Mega · 2023–2025
−15%
Operational costs cut through a rigorous experimentation framework applied to intervention and capital allocation decisions.
Bank Mega · 2023–2025
06

Writing & talks

🎤
Data Science Academy Camp — COMPFEST 14, University of Indonesia
Speaker · 2022 · Medium · Applied data science for students and early practitioners
💻
Open source work & projects — github.com/givensilalahi
Code, experiments, and ML projects on GitHub

Let's work together

Open to consulting engagements, senior DS/ML roles,
speaking invitations, and advisory work in SEA fintech.

✉gjsilalahi@gmail.com inlinkedin.com/in/giventrisjeremia ghgithub.com/givensilalahi