Compare commits
No commits in common. "a760217f3c6a997197fefef090c5b749916f4965" and "6dfe9dc9e0e48b73af0fc6d8f3d6386623232418" have entirely different histories.
a760217f3c
...
6dfe9dc9e0
57
my_cv.yaml
57
my_cv.yaml
|
|
@ -9,8 +9,8 @@ cv:
|
||||||
|
|
||||||
sections:
|
sections:
|
||||||
summary:
|
summary:
|
||||||
- "後端工程師,具備運用 Azure OpenAI 與 LangChain 獨立完成企業級 RAG 與 Text-to-SQL 微服務從 POC 到可部署原型的完整交付能力,協助企業 CRM 系統實現自然語言跨表單查詢 AI 化。"
|
- "專注於現代化 Python 後端架構與 AI 應用的軟體工程師,具備運用 Azure OpenAI 與 LangChain 獨立建構企業級 RAG 與 Text-to-SQL 微服務原型的實戰經驗。"
|
||||||
- "主導 Semantic Caching 優化將 LLM API 延遲由 ~3200ms 降至 ~30ms(50 CCU 壓測驗證);熟悉 FastAPI、Pydantic v2 與 SQLAlchemy v2 (Async) 非同步架構,以 RLS 多租戶隔離與 AST 解析 SQL Guard 確保高併發下的安全性與效能。"
|
- "熟悉 FastAPI、Pydantic v2 與 SQLAlchemy v2 (Async) 現代化非同步架構,注重系統安全性與高併發處理,能實作 RLS 租戶資料隔離與 AST 解析防護機制,並具備效能瓶頸排查與快取優化能力。"
|
||||||
|
|
||||||
education:
|
education:
|
||||||
- institution: "National Taipei University (國立臺北大學)"
|
- institution: "National Taipei University (國立臺北大學)"
|
||||||
|
|
@ -44,16 +44,6 @@ cv:
|
||||||
- "於畢業後參與 300 小時密集實訓,並獨立考取「iPAS AI 應用規劃師 (機器學習) 中級」證照,建立扎實的機器學習理論基礎並轉軌生產環境級 AI 開發。"
|
- "於畢業後參與 300 小時密集實訓,並獨立考取「iPAS AI 應用規劃師 (機器學習) 中級」證照,建立扎實的機器學習理論基礎並轉軌生產環境級 AI 開發。"
|
||||||
|
|
||||||
projects:
|
projects:
|
||||||
- name: "企業級 LLM Gateway 智能查詢微服務 (llm-bridge)"
|
|
||||||
start_date: "2025-12"
|
|
||||||
end_date: "2026-03"
|
|
||||||
highlights:
|
|
||||||
- "專案背景:以模組化單體 (Modular Monolith) 架構,為 NPO 多租戶 CRM 系統建構無狀態雲原生 AI 服務,讓 VIP 客戶以自然語言跨表單檢索業務數據。"
|
|
||||||
- "四模式智慧路由:透過 LLM 結構化輸出將使用者意圖強制分類為 FILTER / AGGREGATION / RANKING / ANOMALY_DETECTION,並搭配動態規則引擎進行熱更新業務邏輯注入,杜絕 LLM 幻覺。"
|
|
||||||
- "AST 級防護與語意快取:以 sqlglot 進行 SQL 抽象語法樹深度解析,100% 阻斷 DML/DDL;結合 PostgreSQL RLS 強制注入 territory 邊界實現多租戶資料隔離;pgvector HNSW 向量索引支援語意快取,在 50 CCU 壓測下將 LLM 延遲由 ~3200ms 降至 ~30ms。"
|
|
||||||
- "成本治理與 WAF:Shadow Wallet 整數運算防溢位模型精準計算 Token 成本並以 HTTP 402 攔截超額請求;Redis 分散式鎖 (SingleFlight) 防範快取擊穿;L7 動態黑名單限流器毫秒內阻斷高頻惡意請求。"
|
|
||||||
- "技術棧:FastAPI、Pydantic v2、SQLAlchemy v2 (Async)、PostgreSQL (pgvector、FTS)、Redis、LangChain、Azure OpenAI、sqlglot、Podman、uv。"
|
|
||||||
|
|
||||||
- name: "食道語者語音輔助裝置 (AI 聲學特徵分類系統)"
|
- name: "食道語者語音輔助裝置 (AI 聲學特徵分類系統)"
|
||||||
start_date: "2023-09"
|
start_date: "2023-09"
|
||||||
end_date: "2024-06"
|
end_date: "2024-06"
|
||||||
|
|
@ -73,45 +63,4 @@ cv:
|
||||||
details: "Multi-tenancy, Semantic Caching, Rate Limiting, AST Parsing, SQL Injection Defense"
|
details: "Multi-tenancy, Semantic Caching, Rate Limiting, AST Parsing, SQL Injection Defense"
|
||||||
|
|
||||||
design:
|
design:
|
||||||
theme: sb2nov
|
theme: sb2nov
|
||||||
sections:
|
|
||||||
allow_page_break: true
|
|
||||||
show_time_spans_in:
|
|
||||||
- experience
|
|
||||||
templates:
|
|
||||||
footer: '*NAME -- PAGE_NUMBER/TOTAL_PAGES*'
|
|
||||||
top_note: '*LAST_UPDATED CURRENT_DATE*'
|
|
||||||
single_date: MONTH_ABBREVIATION YEAR
|
|
||||||
date_range: START_DATE – END_DATE
|
|
||||||
time_span: HOW_MANY_YEARS YEARS HOW_MANY_MONTHS MONTHS
|
|
||||||
experience_entry:
|
|
||||||
main_column: |-
|
|
||||||
**POSITION**
|
|
||||||
COMPANY
|
|
||||||
SUMMARY
|
|
||||||
HIGHLIGHTS
|
|
||||||
date_and_location_column: |-
|
|
||||||
LOCATION
|
|
||||||
DATE
|
|
||||||
normal_entry:
|
|
||||||
main_column: |-
|
|
||||||
**NAME**
|
|
||||||
HIGHLIGHTS
|
|
||||||
date_and_location_column: |-
|
|
||||||
LOCATION
|
|
||||||
DATE
|
|
||||||
education_entry:
|
|
||||||
main_column: |-
|
|
||||||
**INSTITUTION**
|
|
||||||
AREA
|
|
||||||
HIGHLIGHTS
|
|
||||||
date_and_location_column: |-
|
|
||||||
LOCATION
|
|
||||||
DATE
|
|
||||||
locale:
|
|
||||||
language: mandarin_chinese
|
|
||||||
last_updated: 最後更新於
|
|
||||||
month: 個月
|
|
||||||
months: 個月
|
|
||||||
settings:
|
|
||||||
current_date: "today"
|
|
||||||
Loading…
Reference in New Issue