_@baodev

I build backend and AI systems for real operational use — explicit state, bounded decisions, and failure-aware automation.

Most of my work sits at the intersection of backend systems, AI workflows, and operational reliability. I care about what happens outside the happy path: retries, restarts, stale state, partial failure, authorization boundaries, and whether a system can explain why it made a decision.

My main commercial project: a seller-controlled agent that handles Shopee buyer conversations, grounds replies in shop knowledge, and can operate unattended without giving the model unrestricted authority.

• Failure-aware runtime — local-first MV3 state, durable locks/debounce, idempotent decisions, retry/backoff, stale-send protection, and recovery when the browser service worker disappears
• Hard trust boundaries — the extension owns the Shopee session; the backend owns LLM/RAG and secrets; signed requests connect the two without moving marketplace credentials across the boundary
• Bounded automation — deterministic compliance and buyer-risk gates can bypass the LLM entirely; seller-controlled auto-send and a global kill-switch cap effect authority
• Grounded responses — structured shop knowledge + RAG, validation, telemetry, and observable send decisions rather than opaque end-to-end generation

_{TypeScript · Chrome MV3 · Fastify · IndexedDB/Dexie · LLM + RAG}  ·  Active commercial project  ·  🔒 Private source

An open-source agent skill and research track for turning messy market evidence into bounded marketing decisions — without collapsing research, strategy, platform behavior, commerce state, causality, and copy into one prompt.

• Decision-first runtime — starts from the current job, freezes already-resolved decisions, identifies the open decision, and loads only the knowledge that can materially change it
• Addressable JIT knowledge — a lightweight routing layer maps logical knowledge IDs to the smallest relevant handbook/platform section, with deterministic lookup, integrity checks, and capability-aware fallback
• Evidence and claim discipline — keeps observation, interpretation, causality, attribution, platform signals, product truth, commercial state, and uncertainty distinct instead of promoting weak evidence into stronger claims
• Cross-environment reasoning — shared models cover customer research, positioning, messaging, diagnosis, experimentation, localization, social/content distribution, commerce, product discovery, and agent-mediated shopping across platform-specific modules
• Adversarial evaluation — runtime architecture, routing, conceptual compression, and commerce extensions are reviewed with explicit claim ceilings rather than treating passing tests as proof of universal correctness

_{Agent Skill · Python tooling · JIT Knowledge Routing · Marketing Research · Commerce · Evaluation}  ·  Building in public / research in progress  ·  🌐 MIT-licensed open source

A completed full-stack Vietnamese reading and audio platform built around CDN-backed content delivery, realtime features, background processing, and mobile-first UX.

• Static chapter/audio delivery through object storage + CDN while the API serves metadata and application state
• Redis-backed background jobs for batched counters, statistics, reconciliation, cleanup, rewards, and ranking workloads
• Realtime chat/notifications, PWA support, JWT + Google OAuth, rate limiting, and production deployment tooling
• Separate media/content paths for image, JSON, and audio-heavy workloads instead of pushing everything through the application server

_{Vue 3 · TypeScript · Node.js · MySQL · Redis · Cloudflare R2 · Cloudinary}  ·  Completed system  ·  🌐 Public source

An ongoing research track on state integrity in long-lived AI agents: provenance, authority, temporal correctness, durable execution, and the gap between rich runtime traces and governance-relevant interfaces.

• Separates established foundations from agent-specific hypotheses instead of assuming novelty
• Uses falsification-first gates: weak ideas are narrowed or abandoned rather than rescued after the fact
• Current benchmark work studies whether an already-produced action can still be bound to the exact committed external effect occurrence after crash/recovery boundaries
• Freezes claims, implementation, held-out evaluation, and adjudication rules before observing authoritative results

_{Python · Agent Runtimes · Durable Execution · Provenance · Benchmarking}  ·  Research in progress  ·  🔒 Private until release

A paused product/R&D track for finding products with promising source-market signals on 1688 that still face limited competition on Shopee Vietnam.

• Engine pipeline: source crawling → deduplication → translation → Shopee image/market matching → supply enrichment → evidence-backed scoring and decisions
• Opportunity scoring combines demand, visual gap, estimated margin, local-market gap, competition pressure, data confidence, and explicit risk signals
• Separate SaaS consumes a versioned engine contract and provides auth, subscriptions, billing, opportunity exploration, watch/reserve workflows, and market radar

_{Python · Vision / pHash · Next.js · TypeScript · PostgreSQL · payOS}  ·  Paused  ·  🔒 Private source

These are smaller or more inspectable public artifacts that show additional engineering work beyond the primary projects above.

• BAO.OS / Portfolio — interactive portfolio with command routing, multiple RAG modes/personas, Supabase pgvector retrieval, local deterministic embeddings, an OpenAI-compatible API surface, and server-side redaction boundaries; live LLM/RAG is currently disabled on the public deployment.
• CD1-2 — Linux security monitoring lab combining Suricata, Wazuh, deterministic correlation/risk scoring, local IsolationForest signals, evidence-aware explanations, and controlled response policy.
• Audio Ingest — queue-based YouTube audio ingestion pipeline using Redis workers, FFmpeg, object storage, recovery paths, and database synchronization.

Area	Stack
Languages & Runtime
Backend & Security
Data & Processing
AI & Agents
Infra & Storage

_{Open to backend / AI systems roles where reliability actually matters  ·  reach out}