Studium
AI study workspace - upload documents, get contextual Q&A, and auto-generate flashcards powered by vector search.
Next.jsTypeScriptSupabasepgvector
NicoValerianMarcello.
$whoami →App Dev Intern @ BCA&&ML Research
I build backend systems and applied ML: RAG pipelines, Scopus-indexed research on LLM robustness, and full-stack products. Currently a backend intern at BCA, working on myBCA Bisnis.
JavaSpring BootApache KafkaNext.jsTypeScriptPythonPyTorchpgvectorOracle SQLOpenShiftReactRAG pipelines
About / me
Hi, I'm Nico. I build backend systems and applied ML tools, and I care most about shipping them to real users.
I'm a Computer Science student at BINUS University and an Application Developer Intern at BCA in Jakarta. I completed a Spring 2026 exchange at Korea University. My work spans full-stack engineering and applied ML - from AI-powered study tools and LLM benchmarks for low-resource languages, to full-stack apps in production.
I'm most interested in reliable backend systems and applied ML that solves a concrete problem, and I like owning a feature from the data layer to the interface.
Dean's List · 3.95/4.00 GPA
BINUS University · 2025
Scopus-Indexed Publication
ICAICTA 2025 · first author
International Visiting Student, Seoul
Korea University · Spring 2026
Application Developer Intern
BCA · Jakarta · Jul 2026
Selected / projects
Balatro Coach
AI coaching tool for Balatro - upload screenshots for tactical play advice, joker synergy analysis, and blind survival plans.
PythonReactFastAPIYOLOChromaDB
TangCare
Hyperlocal charity logistics platform for Tangerang - connects donors to foundations via geolocation routing.
Laravel 11LivewireAlpine.jsMySQLLeaflet.js
Robin Benchmark
Evaluation benchmark for LLM robustness under Indonesian code-mixing - blended Indonesian/English with real-world slang and typos.
PythonLLM EvalNLP
AniSwipe
Full-stack anime recommendation system with a swipe-based interface - ML-powered matching using MyAnimeList data.
Next.jsFlaskSupabasePythonMAL API
Nakama
Modern Japanese e-commerce storefront with efficient data fetching, cart management, and seamless client-side navigation.
ReactTanStack QueryReact RouterVite
Work / experience
Jul 2026 - Present
Application Developer Intern
BCA · PT Bank Central Asia Tbk · Jakarta
- Develop backend features for myBCA Bisnis, BCA's business internet banking platform, in Java and Spring Boot.
- Built a monitoring system for stuck bulk-transfer payments, with checkpoints at each key processing stage.
- Built a Kafka-based cache for real-time transaction status updates, and fixed a cache eviction bug in production.
Feb - Jun 2026
Visiting Student
Korea University · Seoul
- Advanced ML coursework; cross-cultural research collaboration.
- One of four students selected from BINUS for the exchange program.
2024 - 2025
L&T Manager
BNCC · Bina Nusantara Computer Club
- Directed 4 staff and 6 IT tutors across 5 tracks (UI/UX, FE, BE, Java, C) - 11 classes, 140+ students.
- Launched the ML track, expanding offerings from 5 to 6 core subjects.
- Organised a company visit program for 50+ students.
Jul 2024 - Jan 2026
Scholarship Mentor
BINUS University
- Guided mentees across 25+ tracked sessions; cohort averaged up to 18% GPA improvement within a year.
2023 - 2024
L&T Activist
BNCC · Bina Nusantara Computer Club
- Developed curriculum modules for the Backend and Java tracks.
- Evaluated 15+ tutor candidates; completed LDP producing a semester-long operational roadmap.
Certifications /& awards
Technical / skills
Frontend
ReactNext.jsTypeScriptTailwind CSSAlpine.jsTanStack Query
Backend
JavaSpring BootApache KafkaOracle SQL / PL/SQLREST APIsPythonFastAPIFlaskLaravelSupabaseMySQL
AI & ML
pgvectorChromaDBYOLORapidOCRBM25
Infra & Tools
GitVercelDockerKubernetes (OpenShift)DigitalOcean
Writing /& research
How I Built a RAG Pipeline That Actually Ships
A production postmortem on embeddings, chunking strategies, and the three weeks I spent fighting pgvector index configs - and what building Studium taught me when I had to do it all over again for Balatro Coach.
#engineeringRead →
Understanding Evolving Sentiment Dynamics in Indonesian Disaster-Related Tweets
Fine-tuning IndoBERT for aspect-sentiment analysis on 4,506 disaster tweets - uncovering how public sentiment toward Indonesian crises shifts over time, and what that tells authorities about unmet needs.
#researchRead →
Let's get
in / touch.
Email is the fastest way to reach me. I'm interested in backend and ML engineering work, and open to research collaborations.