Hi!
I'm Jayson
Senior AI/ML Engineer
About Me
I’m very grateful for this opportunity. I consider creativity and inspiration to flourish when there's a merger of minds from various individuals.
My professional journey began in 2015, initially focusing on software engineering with a keen interest in Python and AI technologies. I quickly progressed, taking on challenging roles and AI projects. My experience in various startups as a CTO focused on leading teams in developing innovative software solutions. I worked in multiple sectors, notably healthcare and automotive, where I've implemented AI-driven systems to enhance operational efficiency.
My technical skills are broad, encompassing Python, Django, and AWS. I'm dedicated to leveraging my AI and software development expertise to drive organizational success in this dynamic field.
I value knowledge-sharing among our community, and my objective is to bring solid expertise in practical, real-world AI/ML systems to help you in your day-to-day work and enhance your creativity and vision in product development.
Ultimately, I want to share with you the endless capabilities you can possess to evolve.
My Skills
I am a Senior AI Engineer, with 10 years of experience in the Software Development industry. After graduating with honors in computer science, I’ve been working on architecting, designing, and implementing top-of-the-line software solutions tailored to the unique needs of businesses. As a lead developer in various startups, I led teams in creating innovative software across the healthcare and automotive sectors, implementing AI systems to boost efficiency I developed and maintained various architectural standards, best practices, and methodologies. Also I evaluated team project and selected appropriate technologies and platforms for the solution.
My Portfolio
A small gallery of recent projects chosen by me. I've done them all together with amazing people from companies around the globe. It's only a tip of iceberg compared to the entire list.
- A real-time streaming pipeline (deployed on AWS) that listens to financial news, cleans & embeds the documents, and loads them to a vector DB.
- A fine-tuning pipeline (deployed as a serverless continuous training) that fine-tunes an LLM on financial data using QLoRA, monitors the experiments using an experiment tracker and saves the best model to a model registry.
- An inference pipeline built in LangChain (deployed as a serverless RESTful API) that loads the fine-tuned LLM from the model registry and answers financial questions using RAG (leveraging the vector DB populated with financial news in real-time).
- Unlock the Secret to Efficient Batch Prediction Pipelines Using Python, a Feature Store and GCS and Orchestrate Everything with Airflow.
- Ensuring Trustworthy ML Systems With Data Validation and Real-Time Monitoring for Quality and Integrity using GE. Model Performance Continuous Monitoring.
- 𝗗𝗲𝗽𝗹𝗼𝘆𝗲𝗱 a scalable & cost-effective 𝗮𝘀𝘆𝗻𝗰 𝗯𝗮𝘁𝗰𝗵 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 on top of AWS ECS & SQS that scales in and out based on the number of jobs, 𝗿𝗲𝗱𝘂𝗰𝗶𝗻𝗴 their 𝗔𝗪𝗦 𝗰𝗼𝘀𝘁𝘀 𝗯𝘆 𝟱𝟮%.
- Architecting and building a real-world LLM system from start to finish - from data collection to deployment.
- Leveraging MLOps best practices, such as experiment trackers, model registries, prompt monitoring, and versioning.
- Splitting the monolithic ML system into 4 Python micro services - Data collection, Feature, training and inference pipeline
- It’s a system that simplifies the processing of ensuring compliance with state and insurance regulations for patient charts in the behavioral health field.
- Implementing advanced chunking approach and route system that dynamically choose between semantic RAG, GraphRAG(Fix Entity Architecture), Text-to-SQL and Elastic Search.
- Finetuning the LayoutLMv2 for Visual Document Quesetion Answering such as identifying the state of checkbox and signature information
- Developed AI agent that support neighbors-chat, simulating like human and detecting emotion.
- Developed fast, and extensible multimodal LLM that can understand text as well as human speech, without the need for a separate Audio Speech Recognition (ASR) stage.
- Implemented Agent accross the healthcare, marketing, and retail industries
- Developed fast, and extensible multimodal LLM that can understand text as well as human speech, without the need for a separate Audio Speech Recognition (ASR) stage.
- Research like AudioLM, SeamlessM4T, Gazelle, SpeechGPTs
- Extended Meta's Llama 3 model with a multimodal projector that converts audio directly into the high-dimensional space used by Llama 3
- Respond much more quickly than systems that combine separate ASR and LLM components
- Create new GPT with multi models(GPT-4, Rasa, Llama 1, Llama 2)
- Integrated with supabase(PostgreSQL-Vector)
- Mobile friendly website, AWS
- Solving GPT token limit for prompt(managing context)
- Built this for leveraging personalized insights, generative chat, and automated workflows to empower workforce.
- PGVector for similarity search, RESTful API Development
- Mobile friendly website, AWS
- Solving for limiting the data to owner's bussiness
- Developed RAG to use LLMs to query user data, transform it, and generate new insights.
- Pinecone for building scalable and efficient similarity search applications.
- IONOS
- Solved the problem gpt has by connecting to these data sources and adding user data to the data LLMs already have.
- Implemented core functionality for supporting text to voice, voice to text, meeting, integrating with Whatsapp, Slack
- Integrated with supabase(PostgreSQL-Vector)
- Mobile friendly website, AWS
- Solving GPT token limit for prompt(managing context)
- Huggingface, blenderbot
- Google cloud, Docker
- Created Web Crawler and package it to Media
- RESTful API Development
- Bookstore - React, Django / AWS
- Mobile friendly bookstore
- Figma to React Components
- Canvas JS Animation
- Learning Management System
What my clients say
"Great! I had the pleasure of working with Jayson on a project. Jayson did an outstanding job and delivered the project ahead of schedule. His expertise in these technologies was evident throughout the project, and they were able to quickly troubleshoot and resolve any issues that arose. In addition to their technical skills, Dusan was also a pleasure to work with. he communicated clearly and promptly, provided regular updates. Overall, I am extremely satisfied with the work would highly recommend him to anyone looking for a skilled developer. I look forward to working with him again and will hire him in the future."
Alex Satava, Co-founder - XnetCo
"Jayson is a really nice guy. Was a pleasure to work with him."
Hadi Reda, CEO & Managing Director - Sneakers123
"Resolved the data and scheduling challenges. Excellent communication during the process. Great job!"
Elie Venezky, President - Biometric Edge Inc
Contact Me
I’m interested in freelance opportunities - especially ambitious or large projects.
However, if you have other request or question, don't hesitate to use the form.