杰雷米查莱特
验证专家 in Engineering
数据科学家, 机器学习 Engineer, and Software Developer
Jeremie is a full-stack 数据科学家 focused on natural language processing and tabular data. 他在IT行业有十年的经验, 为英国政府工作, 大型企业, and startups while taking multiple hats as a back-end developer, DevOps, 数据科学家, 创业公司的技术联合创始人. 他领导了两个ML和心理健康社区, 最后但并非最不重要的是, he is an avid reader and lifelong learner who did many retreats.
Portfolio
Experience
Availability
首选的环境
PyCharm, 亚马逊网络服务(AWS), Jupyter笔记本, Slack, Git, Trello, Python, Notion, Linux
最神奇的...
...project I've done is just my last; I built a NLP framework on few-shot learning, 建立主动学习, 并提供了进行偏见分析的工具.
工作经验
自由职业全栈数据科学家
Sprout.ai
- Ran experiments and built a framework to classify text with tiny labeled datasets; few-shot learning using natural language inference.
- Provided a series of tools to better understand the company data and models, 比如主题建模, 模型explainability, 偏差分析, 异常值检测.
- Set up a complete active learning workflow using Label Studio and dedicated back ends.
- Packaged products such as NLI framework and active learning workflow using Docker containers.
- Audited the company's ML platform on MLOps and identified next improvements to focus on.
- Set up a development environment for Kubeflow with Kubernetes and migrated the NLI framework to the Kubeflow pipeline.
- Conducted research on federated learning and suggested a series of bespoke and open-source solutions.
全栈数据科学家
英国创业
- Evaluated and compared a series of MLOps platform solutions like AWS SageMaker, Databricks, Kubeflow, and Cnvrg.
- Designed and proposed multiple service architectures to implement MLOps.
- 使用DVC设置mlop, MLFlow, 和SageMaker来跟踪实验, 火车模型, save, 然后部署它们.
- Performed topic modeling and built a series of ensembles on text classification to identify stress and stressors in social media: a multi-modal deep learning classifier combining text and metadata. 这是数据科学训练营的一部分.
- 生成综合数据来训练和评估模型.
- Ran profiling to provide 8x improvement in inference time on a classifier.
- 为Updraft和一家秘密创业公司做自由撰稿人.
数据科学家
数字、文化、媒体和体育部(DCMS)
- Performed a literature review of state of the art in job offers classification.
- 使用spaCy相似性API构建模型, comparing job offer descriptions and titles to UK Standard Industrial Classification (UK SIC) descriptions.
- Built scripts to run the model on the whole collection (more than one million job offers) and run daily on new job offers.
AI研究员和高级开发人员
国家档案馆
- Interviewed and assessed five suppliers on their solutions and reports from off-the-shelf record management products to bespoke solutions using fully customized models or AI APIs. 参与第三方技术评估项目.
- Wrote a 50+ page report on NLP techniques and tools to select for permanent preservation records held by government departments to be shared with multiple audiences, 包括政府决策者, archivists, 数据科学家.
- Delivered a new release on DROID, an open-source project. 执行或审核60多个拉取请求.
- 在DROID上管理GitHub社区, 响应用户查询, 审查和合并拉取请求.
- Increased project transparency and improved project prioritizing.
- Advocated for the improvement of remote work in my department and offered guidance and support during the COVID-19 lockdown.
首席技术官兼联合创始人
Trackener
- Delivered an IoT product to the market to allow horse owners to look after their horses when they have chronic or acute health issues, 让他们的马保持健康和快乐, 找到内心的平静.
- 领导我们产品的软件部分, 包括移动, web app, 以及后端开发, 服务器管理, 数据科学.
- 负责技术团队的项目管理, including two hardware and software engineers and part-time contractors.
高级Java开发人员
国家档案馆
- 与一名研究员和一名数据专家密切合作, 接管了一个连接文件的原型, and designed and 实现 a set of applications to link collections, 评估它们并发布它们.
- Designed, 实现, and deployed to live and maintained a set of back-end applications dedicated to the categorization of 20+ million records of the national archives for their end-user website Discovery, 使用Lucene, then Solr.
- 管理运行我的应用程序的服务器.
- Installed continuous integration platforms such as Jenkins, Nexus, and SonarQube.
- Created an ML prototype to classify documents based on similarity with Lucene.
- Ran a series of technical presentations to my department.
软件工程师
Atos的《欧博体育app下载》
- 参与开发, 项目管理, 生产支持 of a mediation platform for a high-visibility project for Orange.
- Contributed to multiple projects of varying size on the back ends of the leading French telecom company Orange, 包括项目管理和客户关系, 功能和技术设计, 实现, 生产支持.
- Helped develop a banking application dedicated to mandating management in the SEPA norm for BPCE.
- 监督一个位于印度的离岸开发团队, 包括支持, validation, 和监控.
Experience
马匹护理的物联网产品
我负责我们产品的软件:手机, web app, 以及后端开发, 服务器管理, 数据科学.
Education
计算机科学硕士(MSc)学位
斯塔福德郡大学-斯塔福德，英国
Master of Engineering (MEng) Degree in General Engineering
Ecole Catholique d'Arts et Métiers de Lyon - Lyon, France
认证
Startupbootcamp物联网连接设备
Startupbootcamp
Skills
库/ api
Scikit-learn, Pandas, NumPy, Matplotlib, PyTorch, SpaCy, Natural Language Toolkit (NLTK)
Tools
Slack, Git, PyCharm, Trello, YouTrack, G Suite, Amazon弹性容器服务(Amazon ECS), Seaborn, Notion, Gensim, 亚马逊SageMaker, JProfiler, Apache Maven, NGINX, CircleCI, MQTT, Plotly
Languages
Python, Java, HTML, JavaScript, SQL, CSS, PHP, Groovy
Platforms
Jupyter笔记本, Linux, Ubuntu, 亚马逊网络服务(AWS), Docker, Dataiku, Amazon EC2, Kubeflow, Azure
Paradigms
面向服务的架构(SOA), DevOps, 面向对象设计(OOD), MapReduce
Storage
MongoDB, Neo4j, Amazon S3 (AWS S3)， MySQL
Frameworks
反应本地, Spring Boot, Spring, Flask
Other
计算机科学, 软件工程, 软件架构, 机器学习, 自然语言处理(NLP), 性能调优, GPT, 生成预训练变压器(GPT), 拥抱的脸, 深度学习, 工程数据, 交互设计(IxD), GloVe, MLflow, 数据版本控制, 谷歌BigQuery, 时间序列分析, 主题建模, 主动学习, 机器学习操作(MLOps), 创业, Profiling, 创业加速器
如何使用Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
分享你的需求
选择你的才能
开始你的无风险人才试验
对顶尖人才的需求很大.
开始招聘