Bazzi LogoBAZZI.SYSv2026.0
HEADER.SYS[EXPANDED]
[0,560]
[900,560]
[0,0]
[900,0]
WELCOME TO MY SPACE

Bazzi Tr — Version 2026.0

SCROLL
SECTION_01 — ABOUT
[0,0][600,0]
SPECIMEN_DATA

Trương Đình Khoa - Bazzi Tr

Data & AI Engineer

Architecting intelligent data systems at the intersection of retrieval-augmented generation and scalable ML infrastructure. Specializing in Energy-Based Retrieval (EB-RAG), real-time data pipelines, and autonomous racing intelligence — building systems that think, adapt, and scale.

Building systems that sit at the edge of what retrieval, generation, and real-time inference can achieve — with a focus on architectural correctness, observable behavior, and production-grade reliability.

SYSTEM NOMINAL — BUILDING 2026.0
[900,0][900,600]
SYSTEM_SPECSDATASHEET_v2
DESIGNATION
Trương Đình Khoa - Bazzi Tr
[00]
ROLE
Data & AI Engineer
[01]
FOCUS
Hybrid System Architect & AI Researcher
[02]
LOCATION
Remote / Global
[03]
STATUS
Available for Projects
[04]
SPECIMEN VERIFIED✓ ACTIVE
SECTION_02 — LAB SPECIFICATION
[0,0][1200,0][0,800][1200,800]
LAB_SPEC_001

EB-RAG

Energy-Based Retrieval-Augmented Generation

STATUS

ACTIVE

RESEARCH IN PROGRESS

ARCHITECTURE_OVERVIEW

Retriever (Hybrid search)Energy ScorerContext IntegratorLLM GeneratorEnergy Constraint LayerOutput

EB-RAG is a hybrid architecture combining the expressiveness of Energy-Based Models (EBMs) with the factual grounding of Retrieval-Augmented Generation. Instead of relying solely on likelihood-based scoring, the system uses an energy function to evaluate the compatibility between retrieved context and generated responses — enabling more nuanced, controllable, and grounded AI outputs.

[00]

MRR@10

0.5676

[01]

NDCG@10

0.655

[02]

RECALL@10

1.0

[03]

LATENCY MEAN

132.47ms

[04]

LATENCY_P50

125.66ms

[05]

LATENCY_P95

196.43ms

[06]

THROUGHPUT_QBS

7.55

TECH_STACK_COMPONENTS

PyTorchNumpyDatasetFastAPITransformersHuggingFace
SECTION_03 — MODULES

PROJECT_INVENTORY

Technical Modules

[0,0][400,200]
MOD_001
ACTIVE

RAGEve

RAG System Module — v1.0.0

01

RAGEve is a local-first RAG (Retrieval-Augmented Generation) platform built for developers and teams who want the power of RAG workflows without depending on external cloud services. Everything runs on your own machine — no API keys, no data leaves your network.

MODULE_SPECIFICATIONS
THROUGHPUT1,200 req/min
LATENCY_P99142ms
ACCURACY96.1%
DockerOllamaQdrantLLMsFastAPIPostgreSQL
[400,200][800,400]
MOD_002
ACTIVE

Streamflow

Data Pipeline Module — v1.8

02

Use the SSI API to process tick-by-tick stock price data in Vietnam. Live price quotes, order books, candlestick charts, and market indices for Vietnam's HOSE/HNX/UPCOM exchanges, with a machine-learning-ready feature store underneath.

MODULE_SPECIFICATIONS
THROUGHPUT50M events/s
DATA WAREHOUSE9 DIM / 3 FACT
ACCURACY99.97%
Apache KafkaSparkPythonApache FlinkDocker
[800,400][1200,600]
MOD_003
ACTIVE

AutoRace 2025

3rd Place Finish at Auto Race — v3.0

03

End-to-end autonomous racing system combining computer vision, trajectory optimization, and reinforcement learning for real-time decision making at the edge. It was an intense day of competition where we put our autonomous driving algorithms to the ultimate test.

MODULE_SPECIFICATIONS
THROUGHPUT60 FPS
LATENCY<16ms
ACCURACY98.4%
PyTorchOpenCVDonkeyCarCUDANVIDIA Jetson
SECTION_04 — SYSTEM COMPONENTS

TECH_STACK_OVERVIEW

System Components

[0,0][200,100]
Languages[4 UNITS]
PythonTypeScriptSQLJava
[200,100][400,200]
ML / AI[5 UNITS]
PyTorchTransformersLangChainHuggingFaceRAG
[400,200][600,300]
Data[4 UNITS]
Apache KafkaApache SparkApache FlinkAirflow
[600,300][800,400]
Infrastructure[5 UNITS]
KubernetesDockerAWSAzureLinux
[800,400][1000,500]
Web / Apps[4 UNITS]
Next.jsReactFastAPIVibecode...
SECTION_05 — CONTACT
[0,0][600,800]
INITIATE_CONNECTION

INITIATE CONTACT

Open to collaboration, research inquiries, and system design discussions.

SYSTEM_STATUS

INBOXMONITORED
RESPONSE_TIME< 48H
STATUSAVAILABLE
[1200,0][0,800]
TRANSMIT_MESSAGE

Fill out the form below to initiate a connection request.