Marlon Cajamarca

A passionate seasoned Machine Learning Research Engineer with over half a decade of professional expertise in Machine Learning R&D. I specialize predominantly in the domain of Natural Language Processing, with a marked emphasis on Large Language Models (LLMs). My work traverses a broad spectrum, ranging from Computer Vision to Forecasting, Recommender Systems, and Graph Neural Networks.
I am committed to close the gap between academic research and its industrial application. Adept at infusing groundbreaking research insights into tangible AI solutions, my strength lies in applied R&D geared towards crafting Enterprise-ready Generative-AI applications. Currently I focus entirely on areas encompassing Transformer-based Large Language Models and Computer Vision applications.
I have acquired practical knowledge of ML and LLM operational tools like AWS Sagemaker, ML-Flow, W&B, OpenLLM, H2O-Studio, and Aim, ensuring the streamlined management of ML/LLM models throughout their life cycle (continuous training, evaluation, and monitoring). I've gained experience in containerizing Deep Learning applications using tools like Docker and serving them using Hugging Face Inference Endpoints, AWS Sagemaker Batch Inference Jobs, and other LLM-specialized serving frameworks like vLLM, Ray, and DeepSpeed.
For application development, I tend to harness the capabilities of FastAPI to craft intuitive APIs and employ Gradio, ChainLit, and Streamlit for GUI application endeavours. I exhibit proficiency in designing semi-supervised human-annotator pipelines that are instrumental in refining Large Language and Object Detection models through fine-tuning.
When it comes to databases, I'm well-acquainted with both SQL & NoSQL variants, including but not limited to Neo4J, MongoDB, Redis, Qdrant, Pinecone, and PostgreSQL.
However, beyond the technicalities, I'm deeply passionate about mentoring and teaching. I find joy in simplifying intricate concepts, fostering vibrant discussions, debates, and collective growth within teams.
Lastly, effective communication is a cornerstone of my professional profile. Fluent in English and Spanish, I take pride in elucidating technical blueprints, roadmaps, and intricate design choices in an easily digestible manner for both technical and non-technical stakeholders alike.

I am committed to close the gap between academic research and its industrial application. Adept at infusing groundbreaking research insights into tangible AI solutions, my strength lies in applied R&D geared towards crafting Enterprise-ready Generative-AI and traditional ML applications. Currently I focus entirely on areas encompassing Transformer-based Large Language Models and Computer Vision applications..

BSc Mechatronics Engineering: Mechatronics Engineering bachelor degree with strong emphasis in intelligent systems ans AI.Final grade: Laureate Undergraduate Honors
Thesis: Geometric Approach for Dynamic Modeling of non-redundant, simple and open chain robot Manipulators