You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: data/en/sections/about.yaml
+1-1Lines changed: 1 addition & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -25,7 +25,7 @@ I hold a **BS** from **UCSB** with a double major in **Mathematics** and **Stati
25
25
With a foundation in mathematical and statistical modeling, my research interests lie in leveraging machine learning and deep learning models for innovative solutions.
26
26
I am experienced in modeling and extracting meaningful insights from complex, large-scale datasets across a diverse range of domains, including business, finance, sports, environmental science, and game industry, among others.
27
27
<br/><br/>
28
-
I am actively looking for 2025 summer internship opportunities in Machine Learning, Data Science, or Software Engineering roles!
28
+
I am actively looking for 2026 New Graduate opportunities in Machine Learning, Software Engineering, or Quantitative Research roles!
29
29
'
30
30
31
31
# summary: 'I am a recent graduate from **UCSB** with a double major in **Mathematics** and **Statistics & Data Science**. With a foundation in mathematical and statistical modeling, my research interests lie in leveraging machine learning and deep learning models for innovative solutions. I am experienced in modeling and extracting meaningful insights from complex, large-scale datasets across a diverse range of domains, including business, finance, sports, and satellite imagery, among others.
Copy file name to clipboardExpand all lines: data/en/sections/experiences.yaml
+37-16Lines changed: 37 additions & 16 deletions
Original file line number
Diff line number
Diff line change
@@ -10,6 +10,26 @@ section:
10
10
11
11
# Your experiences
12
12
experiences:
13
+
- company:
14
+
name: Abama Private Fund Investment Management Co., Ltd.
15
+
location: Shanghai, China
16
+
logo: /images/sections/experiences/abama.png
17
+
# darklogo: /images/sections/experiences/
18
+
# company overview
19
+
# overview: "The world’s leading platform for creating and operating interactive, real-time 3D (RT3D) content. "
20
+
21
+
positions:
22
+
- designation: Quantitative Research Intern
23
+
start: July. 2025
24
+
# don't provide end date if you are currently working there. It will be replaced by "Present"
25
+
end: Aug. 2025
26
+
# give some points about what was your responsibilities at the company.
27
+
responsibilities:
28
+
- Designed a parallelized data pipeline to transform raw event-level financial data into 300+ engineered featuresto feed high-frequency trading ML models, cutting data preparation from 8 h to 15 min.
29
+
- Delivered a live-updating factor store database powering large-scale model training and backtests; increased experiment throughput by ∼ 10× and enabled apples-to-apples model comparison during strategy reviews.
30
+
- Fine-tuned Transformers for trading signal prediction, improving annualized return by 3% in backtests.
31
+
- Built an automated feature-discovery framework that iteratively surfaces noise-robust, signal-strengthening factors, accelerating model iteration and reducing manual screening.
32
+
13
33
- company:
14
34
name: Unity
15
35
url: "https://unity.com/"
@@ -20,15 +40,15 @@ experiences:
20
40
# overview: "The world’s leading platform for creating and operating interactive, real-time 3D (RT3D) content. "
21
41
22
42
positions:
23
-
- designation: R&D Sofeware Engineering Intern
43
+
- designation: R&D Software Engineering Intern
24
44
start: May. 2024
25
45
# don't provide end date if you are currently working there. It will be replaced by "Present"
26
46
end: Aug. 2024
27
47
# give some points about what was your responsibilities at the company.
28
48
responsibilities:
29
-
- Scaled the RAG database for Unity’s AI assistant, Muse Chat, by developing a workflow that systematically gathers inputs from various Unity forums and enhances data quality with an LLM pipeline.
30
-
- Created MuseBench, a system using LLM-as-a-judge along with a benchmark dataset to evaluate the AI agents, achieving a 95% reduction in time required for assessment processes while minimizing human labor.
31
-
- Developed a pipeline with a locally deployed LLM to extract the key error messages from Unity Cloud Build log files that typically exceed 100k lines.
49
+
- Scaled the RAG knowledge base for Unity’s AI assistant via community and forum ingestion and LLM-based quality filtering; expanded answer coverage for Editor/API issues by ∼ 10×, improving first-response resolution.
50
+
- Built MuseBench, an evaluation platform that pairs an LLM-as-a-judge with a curated benchmark of Q&A pairs, enabling consistent scoring of RAG results across different model versions and cutting evaluation time by 95%.
51
+
- Deployed a local-LLM pipeline to parse 100k+-line Unity Cloud Build logs and extract root-cause signals; reduced triage from hours to minutes and accelerated time-to-fix for recurring build failures.
- Trained and fine-tuned various CNN and transformer models for Image Super-Resolution
74
-
- Led a comprehensive ablation study to evaluate and refine the proposed architecture
75
-
76
-
- designation: Summer Intern
77
-
start: Jun. 2022
78
-
end: Aug. 2022
79
-
responsibilities:
80
-
- Conducted literature reviews on Image Super-Resolution deep learning models.
89
+
- designation: Undergraduate Research Assistant
90
+
start: Aug. 2022
91
+
end: Jun. 2023
92
+
responsibilities:
93
+
- "Co-authored the ACM MM '24 paper :'[Compacter: A Lightweight Transformer for Image Restoration](https://openreview.net/pdf?id=gvrDYlQXxw)', achieving state-of-the-art PSNR performance across different Image Restoration tasks with ~ 50% - 65% fewer parameters."
94
+
- Co-designed Compact Adaptive Self-Attention, enabling omnidirectional spatial–channel information flow through cross-modulation of global context to strengthen long-range dependencies while preserving local detail.
95
+
- Proposed a Dual Selective Gated Module that dynamically injects global context into each pixel for contextadaptive aggregation, amplifying informative features and suppressing noise.
96
+
- Built the PyTorch training/benchmarking pipeline and ran ablation experiments, enabling reproducible results and efficient comparison to baselines.
97
+
- designation: Summer Intern
98
+
start: Jun. 2022
99
+
end: Aug. 2022
100
+
responsibilities:
101
+
- Conducted literature reviews on Image Super-Resolution deep learning models.
0 commit comments