# 伟琪 唐

伟琪 唐 is a remote worker on Himalayas, AI Applications Engineer at Alibaba International, based in China.

Source: https://himalayas.app/@0009687

```text
AI Applications Engineer at Alibaba International building gold benchmarks and improving LLM agent task success.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | China |
| Authorized to work in | China |
| Salary expectations | 45k+ USD |
| Job categories | AI Engineer, Data Engineer, Machine Learning Engineer |
| Skills | AI Agent Evaluation, Dataset Curation, Trajectory Filtering, Golden Set Benchmarking, Judge Calibration, Text To SQL, SFT, GRPO, Hybrid Retrieval, Reranking, Data Quality Checks, Python, SQL, DuckDB, PyTorch |

## About

```text
At Alibaba International, I built a 300-case gold benchmark from historical merchant root-cause analyses and calibrated an LLM judge against human ratings, reaching 92% agreement.

I mined agent traces and tool calls for regression cases, helping raise offline pass rates across two workflows from about 71% to 84%. I also collected and filtered teacher trajectories for action-only SFT and designed verifiable rewards for online GRPO; strict task success on held-out shopping tasks rose from about 53% to 67%.

At Kuaishou, I built a stratified text-to-SQL evaluation set from 260 real analytics questions and added data-quality checks. Correct table selection rose from 75% to 88%, while SQL execution success increased from 72% to 86%.

For an independent DuckDB project, I built a data pipeline repair benchmark using NYC taxi-trip data across two monthly Parquet releases. Its automated verifier passed the reference SQL repair and detected failures involving dropped rows, NULL handling, and mismatched monthly fee totals.
```

## Experience

### AI Applications Engineer (current)

Alibaba International. Apr 2026 - Present

```text
Built a 300-case gold benchmark for LLM agent evaluation and calibrated an LLM judge with 100 human-labeled examples, achieving 92% agreement. Mined agent traces for regression cases, improving offline pass rate from 71% to 84%, and collected teacher trajectories for SFT and GRPO, raising task success from 53% to 67%.
```

Skills: LLM-Evaluation, Benchmark-Curation, Trajectory-Filtering, SFT, GRPO, Python, SQL, Agentic-AI

### AI Data Engineering

Kuaishou. May 2025 - Aug 2025

```text
Built a stratified text-to-SQL evaluation set from 260 real analytics questions, improving table selection from 75% to 88% and SQL execution success from 72% to 86%. Curated 1,800+ knowledge chunks and 420 SQL examples, boosting top-5 recall from 77% to 91% with hybrid retrieval and reranking.
```

Skills: Text-To-SQL, Dataset-Curation, Evaluation, SQL, Python, Retrieval, Reranking

## Education

### The Hong Kong University of Science and Technology

MPhil, Data Science and Analytics

```text
Pursuing an MPhil in Data Science and Analytics, supported by the Red Bird Scholarship.
```

### Jilin University

Bachelor's degree, Finance and Accounting

```text
Earned a Bachelor's degree in Finance and Accounting.
```

## Tech stack

- Python
- PyTorch
- SQL
- DuckDB
- LangGraph

To contact 伟琪, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [qingqing he](https://himalayas.app/@qingqinghe): Senior Data Scientist, Tencent
- [junluo yang](https://himalayas.app/@junluoyang): 算法专家 智能聊天系统, 成都美尔贝科技有限公司
- [yimin Lin](https://himalayas.app/@yiminlin): AI Applications Project Lead, Independent Project
- [Zhang Jiaxin](https://himalayas.app/@JiaxinZhang): LLM Data Engineering Lead, AI Assisted Web Fiction Content Data Platform
- [JB Bryant](https://himalayas.app/@jbbryant): Applied AI Practitioner, Kaggle
- [Boxuan Tian](https://himalayas.app/@boxuantian): Senior Model Evaluation Expert, Xiaotian Ruishi (Alibaba Group)
- [muding mao](https://himalayas.app/@mudingmao): Operations Development Engineer, Huawei Technologies Co., Ltd.
- [Spike User](https://himalayas.app/@spikeuser): AI Engineer, Data Scientist, Team Leader, MDPI
- [Spike Xiong](https://himalayas.app/@spikexiong): AI Engineer, Data Scientist, Team Leader, MDPI
- [千雯 丛](https://himalayas.app/@0009244): Freelance AI Evaluation Task Designer, TalentsAI
