Shengheng Yan

Email: sh.yan.shengheng@gmail.com
GitHub: https://github.com/timbrist

Experience

Research Assistant

Aalto University
Espoo, Finland
June 2024 - August 2024

Developed a CARLA autonomous driving agent using large language models like LLama2, and GPT4o.
Developed a natural language narrations and reasoning system for driving dataset from BDD-X-dataset, using GPT4o
Implement and finetuning the state-of-art method on BDD-X-dataset on videollava
Design experiments and evaluate results on metrics like BLUE, ROUGE, and CIDEr.

Summer Intern

University of Jyväskylä
Jyväskylä, Finland
April 2023 - August 2023

Developed an interpreter of Multi-objective optimization problems for group framework DESDEO.
Implement a web interface using Typescript, and CortexJS library. It allows users to type multi-objective optimization problems in mathematic symbols and use DESDEO features.
Wrote unit tests to ensure all the multi-objective optimization problems work correctly.

Game Developer

Shanghai CloudToad Network Technology Co., Ltd.
Shanghai, China
August 2021 - July 2022

Developed scout, equipment system on SLG mobile game using Unity, Lua, and C++.
Developed a data error-checking game tool(using VB) for game designers.
Designed protobuf data format for back-end and front-end.
Participated in code reviews and contributed to Git and SVN management.
Bug-fixing and Documentation.

Software Engineer

TIANMA MICROELECTRONICS CO., LTD.
Shanghai, China
July 2020 - August 2021

Developed hibernate and wake-up, backlight dimming, transmission of touch data, and power management on automotive display/screen.
Debugged de/serializer MAX9278/MAX9282, Synaptics TB7800, and ARM Cortex M0/M4 using oscilloscope, TP-Link, Keil C.
Familiar with ISO26262 and MISRA C/C++
Cloudified documents and specifications that are shared among SE, EE, and QA departments.
Fixed critical bug on the IC board.

Software Engineer Intern

Shanghai WenShang Technology CO., LTD
Shanghai, China
July 2019 - December 2019

Implement image algorithms to detect flaws in automotive components using C++, OpenCV, and Tensorflow CNN.
Optimizing the time complexity of the company image analyzing program.
Writing data analysis Python scripts to track detection data and parameters.

Education

Master of Science in Cognitive Computing

University of Jyväskylä, Finland
2022-2024
Thesis: Autonomous Driving Systems with Large Language Models: A Comparative Study of Interpretability and Motion Planning

Investigated the integration of large language models into autonomous driving systems
Implemented autonomous agent using CARLA simulator
Implemented state-of-art models: TransFuser and LMDrive
Corroborated with Aalto University

Bachelor of Engineering in Software Engineering

Shanghai Institute of Technology, China
2016-2020
Thesis: 3D reconstruction with Structure Light.

A non-contact 3D surface measurement technique using C++, OpenCV and a 3D reconstruction algorithm. The code is available as open source on github.

Projects

Hikvision-qt

Description: A Qt interface for Hikvison camera MV-CA013-21UM
Achievements: 40+ stars on github.
Technologies Used:

Home Service Robot

Description: Robot functions include Speech Recognition, Computer Vision and Navigation.
Achievements: My team won 6 prizes at RoboCup (China Region), including 3 The First Prizes, and 9 prizes at China Service Robot Competition.
Technologies Used: