TRANSFORM = IDENTITY

det = 1.00

PORTFOLIO / 2026Singapore · OPEN TO WORK

Backend / AI Engineer

BASILBOH

I build backend systems that stay fast under load and AI pipelines that stay honest under scrutiny.

VIEW WORK GET IN TOUCH

↓ SCROLLî = BASIS-X · ĵ = BASIS-Y · LIVE 2×2 MAP

/ LATENCY · LIVE

ABOUT

WHAT I OBSESS OVER

I optimise the layers most people never look at.

I'm Basil, a backend and AI engineer. I geek out on the parts of a system that decide whether it survives real load: caching, connection pooling, batch processing, async, and heavy query optimisation.

The other half of my brain lives in AI infrastructure. I build RAG pipelines, evals and agent harnesses that are measurable and debuggable instead of held together by vibes.

I'm a Computer Science (Honours) student at NUS, currently interning on the FX Technology team at Singapore Exchange (SGX). Most of what I build starts as a hackathon hack and grows from there.

/ SPEC SHEET

ROLE: Backend / AI Engineer
BASE: Singapore
FOCUS: SYSTEMS · AI · STATS
EDU: NUS CS (HONS) · '27
STATUS: OPEN TO WORK

01AXIS

HIGH-PERFORMANCE SYSTEMS

Heavy query optimisation, caching, connection pooling, batch processing and async I/O. Pooled, pipelined paths and the cold tail latencies nobody profiles until it's on fire.

PostgreSQLRedisFastAPIGoDocker

02AXIS

APPLIED AI / LLM INFRA

RAG pipelines, agent harnesses, tool-calling orchestration and evals that retrieve the right context, not the most context.

PythonOpenAIGemini 3 PropgvectorVertex AI

03AXIS

STATISTICS & MODELLING

Treating latency, recall and cost as distributions, not averages. Load modelling and measurement that turn noisy telemetry into decisions you can defend.

NumPyPyTorchscikit-learnDuckDB

/ THROUGHPUT · STREAM

PROJECTS

13 BUILDS / SYSTEMS + AI

P-01

AI / Analytics2026

SPATIALREHAB

A visionOS companion for early-stage dementia, built on one clinical rule: never let the person feel they got it wrong. Eight short games take a silent cognitive baseline, then the daily exercises rehearse what dementia takes first — making kopi one step at a time, mahjong by the real Singapore rules, and the route home studied on a holographic tabletop before being walked at life size through a neighbourhood meshed from OpenStreetMap. The caregiver dashboard holds the only numbers in the app: one trend line per game, never a grade.

visionOSSwiftUIRealityKitSwift ChartsOpenStreetMap

THE TABLETOP — THE ROUTE HOME

OPEN ↗

P-02

Backend / Security2026

AIRLOCK

A safety gate that stands in front of every package an AI agent installs. Each candidate is detonated in a throwaway Daytona sandbox seeded with honeytoken credentials, read line-by-line by a code model on a rented GPU, and reputation-checked against the live web — then a judge weighs all of it and blocks or passes, in plain language the agent understands.

PythonDaytonaQwen 3.5Nosana GPUDoublewordOxylabs

Agent terminal session where Airlock intercepts a pip install of python-pillow and returns a BLOCKED verdict.

BLOCKED — python-pillow

OPEN ↗

P-03

Mobile / Backend2026

BIOHACKK JOURNEY APP

A full-stack mobile health platform: Singpass and Google OAuth, realtime bidirectional sync across four-plus biometric sources — Apple HealthKit, Omron, smart scales and Visbody body scans — and a phase-aware daily protocol with food, sleep, mood and vitals journaling on top. The Supabase and GCP pipelines behind it ingest 10,000+ readings a day; reworking the schemas and query paths took API responses from 3s to 500ms.

React NativeExpoTypeScriptSupabaseGCPNestJSHealthKit

The nutrition detail screen: a photographed meal identified as strawberry kakigori with creme brulee topping, a health score, and a per-ingredient sodium, salt and energy breakdown.

The quests screen listing daily and weekly missions — 10K steps, mindful moment, workout streak, hydration hero — each with XP and coin rewards and a progress bar.

LIVE — DASHBOARD, PHASES, BODY SCAN

OPEN ↗

P-04

Quant Research2026

PORTFOLIO RISK ANALYTICS

A quant notebook that runs AMZN's 29-year return series through the full risk battery against SPY — Sharpe and Sortino, drawdown curves, beta and alpha, a monthly-return heatmap and a 1,000-path Monte Carlo on the trailing year. A 30% CAGR reads very differently once you plot the drawdown underneath it.

Monte Carlo simulation of 1,000 paths over 252 trading days, with the realised return traced in red through a 95% confidence band.

MONTE CARLO — 1,000 PATHS

PythonquantstatspandasNumPyJupyter

HDB RESALE PRICE MODEL

A linear regression over 287,196 Singapore HDB resale transactions, built as an honest end-to-end scikit-learn pipeline: correlation and multicollinearity screening, leakage checks, imputation kept inside the pipeline, and a held-out test set measured against a mean-predictor baseline rather than against itself.

Scatter plot of resale price against floor area for Singapore HDB flats, with the fitted linear model drawn through it in red.

FIT — RESALE PRICE VS FLOOR AREA

Pythonscikit-learnpandasseabornJupyter

MNIST DIGIT CNN

A convolutional net built from the layer definitions up in PyTorch — two conv blocks with dropout into a 50-unit head, 21,840 parameters total — trained for ten epochs on a CPU across all 60,000 MNIST digits and scored against the 10,000-image held-out set after every one. 92% after the first epoch, 97.46% by the tenth, with a loss curve that floors at 1.46 for a reason worth explaining.

Two MNIST test digits rendered from the notebook — a 7 and a 2 — both classified correctly by the trained network.

PREDICTED — 7 AND 2

PythonPyTorchtorchvisionNumPyJupyter

ETL PIPELINE

In build. A scheduled extract–transform–load pipeline: pull from source APIs, land the raw payloads untouched, then parse, validate and reshape them into warehouse tables a query can trust. Reruns are the design constraint — the same window replayed twice should land the same rows once. Numbers, source and captures go up when it ships.

Schematic of the pipeline: three stages — extract, transform and load — connected left to right, with the sub-steps of each listed underneath.

SCHEMATIC — EXTRACT / TRANSFORM / LOAD

SUB-MS VECTOR SEARCH ENGINE

A from-scratch ANN engine with an HNSW index, SIMD distance kernels, and a lock-free query path. Serves nearest-neighbour lookups over 50M embeddings without breaking a millisecond at p99.

RustSIMDHNSWTokioArrow

DISTRIBUTED RAG ORCHESTRATOR

A DAG-based orchestration layer for retrieval-augmented agents: typed tool calls, speculative retrieval, automatic eval gating, and full request tracing across every hop.

PythonasynciopgvectorTemporalOTel

An interactive tool that turns Postgres EXPLAIN ANALYZE output into a navigable cost graph, surfacing the one missing index hiding behind a sequential scan.

TypeScriptWebGLPostgreSQL

HACKATHONS

7 BUILDS · 3 PODIUMS

H-011ST / 126

APPLE SPATIAL HACK AI

SpatialRehab

AUG 2026 · Singapore

SOURCE

Grand Champion and Best UI/UX — first of 126 participants. A visionOS cognitive rehabilitation tool for early-stage dementia, built on errorless-learning principles: eight baseline games score memory, attention and executive function silently, then patients study a miniature neighbourhood procedurally meshed from OpenStreetMap and re-walk the route at life size in passthrough. Caregivers get trend charts and a clock-drawing gallery tracking decline over sessions.

visionOSSwiftUIRealityKitSwift ChartsOpenStreetMap

H-021ST / 284

DAYTONA HACKSPRINT

Airlock

JUL 2026 · Singapore

Overall Grand Winner. A security gate for AI agents that blocks supply-chain malware before install — detonating each package in a disposable Daytona sandbox seeded with honeypot credentials, fused with Qwen 3.5 static analysis on a Nosana GPU, Doubleword embedding matches against known malware and live Oxylabs reputation intel.

DaytonaClaude Code HooksNosana GPUDoublewordOxylabs

H-032ND PLACE

RALPHTHON @SG

SnapRoom

MAY 2026 · Singapore

Turned a single room image into a walkable 3D world — Gaussian-splat environment generation, AI object meshes, a semantic-label layer and a mobile AR viewer.

Three.jsWorld LabsFal AIExpo

H-04FEATURED BUILD

AI ENGINEER SINGAPORE

DataForge

MAY 2026 · Singapore

An end-to-end AI engineering prototype — agents, retrieval and structured-output evaluation — built alongside OpenAI, Cursor, Vercel, Google DeepMind and ElevenLabs.

Next.jsConvexOpenAIFal AI

H-05LIVE AI DEMO

GRABMAPS

Personality Map

APR 2026 · Singapore

A tap-a-place experience on Grab Maps where archetype characters deliver review-flavoured monologues in ElevenLabs voices, brokered through a key-safe Vercel Serverless BFF.

Grab MapsOpenAIElevenLabsVite

H-06SHIPPED TO CLOUD RUN

NUS HACK&ROLL

Reely

JAN 2026 · Singapore

A GeoGuessr-style game that geolocates uploaded images and videos via Gemini Vision, with FFmpeg scene detection and a Dockerised stack on GCP Cloud Run.

ReactMongoDBGemini VisionGCP

H-07LONG-VIDEO AI

GOOGLE GEMINI

Urban Incident Review

JAN 2026 · Singapore

An AI video analysis platform answering natural-language queries over footage with timestamped, confidence-scored findings using the Gemini Files API and Gemini 3 Pro.

Next.jsGemini 3 ProExpress

/ DISTRIBUTION · SPECTRUM

EXPERIENCE

WHERE I'VE SHIPPED

AUG 2026 — NOW
SINGAPORE EXCHANGE
Singapore
Analyst Intern, FX Technology
- 01Handle deployment, configuration and production reliability for MaxxTrader, a low-latency FX trading platform used by bank, hedge fund and asset manager clients.
- 02Track and analyse MaxxTrader performance metrics spanning latency, throughput and system reliability to surface bottlenecks, partnering with Development, QA and Sales on client-driven technical solutions.
FX TradingLow-Latency SystemsMaxxTraderProduction Reliability
APR 2026 — JUN 2026
TWISS
Singapore
Software Engineer Intern
- 01Building WatchTower, an AI operations layer for F&B — turning RTSP camera feeds into a queryable event stream (queue buildup, table turnover, kitchen congestion) with realtime alerts and natural-language 'ask-your-camera' queries.
- 02Designed a multimodal pipeline fusing YOLO and Vertex AI Vision detections with an OpenAI summarisation layer to emit autonomous operational recommendations.
- 03Shipped a tiered cost-optimisation layer — a cheap vision model for steady-state monitoring, escalating to a stronger multimodal one only on anomaly triggers — cutting daily vision-inference volume from 10M to 1.5M tokens (85% cheaper), on FastAPI + Next.js + Supabase Realtime on Cloud Run.
-85%
INFERENCE COST
1.5M
VISION TOKENS / DAY
FastAPINext.jsVertex AIOpenAISupabase
JAN 2026 — APR 2026
BIOHACKK
Singapore
Software Engineer Intern
- 01Built secure auth with Singpass and Google OAuth using state, nonce and token validation for end-to-end security.
- 02Integrated Apple HealthKit and third-party IoT/biometric SDKs (Omron, smart scales, Visbody) for realtime bidirectional sync across 4+ biometric sources.
- 03Designed Supabase + GCP services and data pipelines ingesting 10,000+ health readings a day, cutting API response times from 3s to 500ms (83% faster) on schemas built for user metrics, streaks, leaderboards and health tracking.
- 04Shipped AI-powered nutritional analysis (OCR + LLMs), push notifications and dynamic data-visualisation dashboards.
-83%
API RESPONSE
10,000+
READINGS / DAY
4+
BIOMETRIC SOURCES
React NativeSupabaseGCPApple HealthKit
MAY 2025 — JUL 2025
DNDTS PTE. LTD.
Singapore
Software Engineer Intern
- 01Built and maintained a full-stack logistical inventory management system tracking SKUs, price points, model numbers and shipments, with React (Vite, Mantine) and a Go backend on SQL + Docker.
- 02Shipped new features and system enhancements; optimised SQL query performance across microservices and internal tools, cutting a core internal query from 5s to 0.28s — 18× faster.
0.28s
CORE QUERY
18×
SPEEDUP
ReactGoSQLDocker
SEP 2024 — SEP 2025
NUS STUDENTS' COMPUTING CLUB
Singapore
Vice-President
- 01Led 100+ members across subcommittees to organise 7 major faculty-wide events, growing total attendance from 3,800 to 5,000 computing students — 32% year-over-year.
- 02Coordinated hackathons, tech talks and workshops while leading sponsorship outreach, growing sponsor funding from $11k+ to $23k+ (109% year-over-year); mentored junior leaders for continuity.
100+
MEMBERS LED
7
MAJOR EVENTS
+32%
ATTENDANCE
+109%
SPONSOR FUNDING
LeadershipEventsSponsorshipsMentorship
AUG 2024 — AUG 2025
NUS FINTECH SOCIETY
Singapore
Business Development Executive
- 01Secured sponsorships and partnerships for large-scale fintech events.
- 02Built strategic collaborations to enhance engagement and funding.
PartnershipsSponsorshipsFintech

/ NETWORK · TOPOLOGY

EDUCATION

FOUNDATIONS + CONTINUOUS

012023 — 2027

NATIONAL UNIVERSITY OF SINGAPORE

B.Comp. (Hons), Computer Science

Bachelor of Computing (Honours), graduating May 2027. Focus on full-stack engineering, distributed systems and applied AI; active across NUS's hackathon and student-leadership communities.

022018 — 2021

/ PHASE · PORTRAIT

07 / CONTACT

LET'S BUILD SOMETHING FAST.

basil.boh001@gmail.com

/ AVAILABILITY

OPEN TO WORK · Singapore

/ ELSEWHERE

BASILBOH

ABOUT

HIGH-PERFORMANCE SYSTEMS

APPLIED AI / LLM INFRA

STATISTICS & MODELLING

PROJECTS

SPATIALREHAB

AIRLOCK

BIOHACKK JOURNEY APP

PORTFOLIO RISK ANALYTICS

HDB RESALE PRICE MODEL

MNIST DIGIT CNN

ETL PIPELINE

SUB-MS VECTOR SEARCH ENGINE

DISTRIBUTED RAG ORCHESTRATOR

CONNECTION-POOL AUTOPILOT

STREAMING FEATURE STORE

LLM EVAL HARNESS

QUERY PLAN VISUALISER

HACKATHONS

APPLE SPATIAL HACK AI

DAYTONA HACKSPRINT

RALPHTHON @SG

AI ENGINEER SINGAPORE

GRABMAPS

NUS HACK&ROLL

GOOGLE GEMINI

EXPERIENCE

SINGAPORE EXCHANGE

TWISS

BIOHACKK

DNDTS PTE. LTD.

NUS STUDENTS' COMPUTING CLUB

NUS FINTECH SOCIETY

EDUCATION

NATIONAL UNIVERSITY OF SINGAPORE

TEMASEK POLYTECHNIC

NOTES

Backend engineering concepts I've been learning about

Designing a ledger you can trust

How to actually win a hackathon

RAG in production: the boring parts that matter

Shipping is a craft, not an event

The index you didn't know you needed

RAG is a retrieval problem, not a prompt problem

Connection pools as queueing systems

Stop reporting averages

LET'S BUILD SOMETHING FAST.