Machine Learning Ml Engineer

Yuji Oshima

AI/ML Engineer

****.************@*****.*** +1-339-***-**** San Francisco, CA, USA

github.com/YujiOshima kaggle.com/yujioshima

PROFILE

Software Developer turning into a full-fledged Machine Learning Engineer. ML Engineer with +7 years experience and a strong background in LLMs, NLP, Gen-AI and Image Processing. Experience in developing and implementing machine learning algorithms to solve complex business problems across various industries, including fin-tech, tourism, healthcare, e-commerce. Currently acquiring MLOps skills. Proficient in programming languages such as Python and C++, JavaScript with expertise in popular machine learning libraries like Pytorch and Tensorflow.

EDUCATION

Benedictine University

Master's Degree in Computer Science

Apr 2011 – Apr 2017 Naperville, USA

•Relevant Coursework: Algorithms and Data

Structures, Machine Learning.

•Research paper: Tensor Decomposition and its

application in Deep Learning, March 2017

•GPA: 3.6

CERTIFICATES

Artificial Intelligence (AI) Certification

Columbia University

Mar 2019

Global Machine Learning Certification

University of the West

July 2021

Certified Prompt Engineer Certification

Blockchain Council

August 2023

PROFESSIONAL EXPERIENCE

Tredence

ML Engineer

Apr 2023 – Jun 2024 CA, USA

•Designed and developed robust and scalable CI/CD/CT pipelines to automate the end-to-end ML workflow, from data ingestion to model deployment using Vertex AI on GCP.

•Automated ETL data pipeline using Apache Airfow with Kafka, Snowflake, Spark and Tableau and enhanced the scalability via some best practices.

•Developed a RAG Chatbot utilizing LCEL and Graph Database and executed Google Gemini on GCP via Vertex AI.

•Monitored the performance of deployed models and pipelines, using tools like Grafana to track metrics and integrated several model test including Invariance Tests and Data Unit Tests.

•Deployed machine learning models into production environments, using containerization tools like Docker and orchestration platforms like Kubernetes to ensure seamless integration.

•Integrated Large Language Model (LLM) capabilities into resume mathing and analysis using Langchain and executed Mixtral large on Azure via Azure ML.

•Integrated a Vector Database such as FAISS into existing systems to enhance data retrieval and storage efficiency.

•Developed an LLM Hotel Recommender using the Hypothetical Document Embeddings (HyDE) approach with Redis and the OpenAI API.(Tourism)

•Designed a wide range of prompts tailored to specific AI tasks and applications to improve model performance.

Mercari

AI Developer

Apr 2019 – Mar 2023 CA, USA

•Implemented the CI/CD pipeline for training-predicting using Google AI Platform + Airflow for Detection the Chargeback transaction in Merpay payment service using machine and conducted cancel detection of transactions based on chat messages with NLP.

•Fine-tuned the LLM with the QLoRa technique for AI Assist in e-commerce and conducted fine-tuning on CLIP(Contrastive Language-Image Pre-Training) to significantly boost the performance of item category and brand prediction,.

•Improved recommendation accuracy using collaborative filtering and vector search engine and applied an LLM-based approach to large-scale Item category classification.

•Developed and maintained an AI-based recommendation system for an e-commerce platform, resulting in an 18% increase in customer engagement and a 12% increase in sales.

•Enhanced ML pipelines for the deferred payment service (Fin-Tech) with manual confirmation and implemented the model management for client-side ML powered by Firebase.

•Conducted neural architecture search(NAC) to automate the multi-modal modeling for prohibited item detection.

•Used Tableau and SQL to redefine and track KPIs surrounding marketing initiatives, and implemented A/B tests to generate 15% more client leads.

•Reduced fraudulent activities by 49% by creating an anomaly detection system for financial transactions to safeguard the company’s assets.

•Collaborated with cross-functional teams to gather and preprocess data, fine-tune models, deploy several models into production environments and power client-side machine learning with TensorFlow Lite.

•Developed a general customer profiling platform to host machine learning models using Google CloudSQL and Google DataStudio.

ShipBob

Backend Developer

Jun 2017 – Mar 2019 IL, USA

•Developed a serverless architecture using AWS Lambda and API Gateway to achieve a cost-effective and scalable backend solution.

•Managed AWS infrastructure, including EC2 instances, S3 buckets, and RDS databases, ensuring high availability and data security.

•Built and maintained back-end server-side applications using Node.js, ensuring efficient communication between the Server and Client-side components.

•Enhanced an e-commerce site's search capabilities by leveraging Algolia to provide users with fast and relevant search results.

•Implemented user authentication and authorization using industry-standard protocols like OAuth2 and Clerk, resulting in improved security and reduced a risk of data breaches.

•Developed a RESTful API using Node.js and Express, enabling the company to expose its services to external users and systems, resulting in improved efficiency and flexibility.

•Leveraged Salesforce CRM to streamline customer relationship management, enhancing sales and marketing processes.

•Implemented Semantic HTML5 markup throughout the project, enhancing accessibility and improving Search Engine Optimization ( SEO ) by providing clear and meaningful structure to web content. SKILLS

Soft Skills

Team Spirit, Creative Thinking,

Time-Management

Big Data Technologie

Hadoop, Spark, Hive, Kafka,

Pandas

Cutting Edge Technology

ZK(Zero Knowledge), IoT,

BlockChain

Programming Language

Python, C++, R, JavaScript,

Rust

Cloud Computing

AWS ML, Google Cloud AI,

Azure AI

Machine Learning Framework

NumPy, TensorFlow, PyTorch,

Keras, Scikit-learn

Data Visualization Tool

Tableau, Power BI, Plotly,

Matplotlib

PROJECTS

Fine-Tuning FLAN-T5 for DeToxified Text Generation

•Description: Mitigated toxicity in the generated text by fine-tuning FLAN-T5 through reinforcement learning by leveraging Meta AI's pre-trained "hate speech" reward model while preserving creativity and fluency

•Tools: NLP, FLAN-T5, PyTorch and Reinforcement Learning Real Time Conversational HR Bot

•Description: Developed a conversational chatbot using RASA for pre-screening interviews and shortlisting candidates. The Rasa framework was implemented with MongoDb for tracking data.

•Tools: RASA, MongoDb, Flask, Javascript, Docker, Google APIs

Contact this candidate