Explore practical techniques for compressing prompts in large language models, enhancing performance and efficiency.
NEU '24, M.S Student, AI/ML Engineer
I'm a dedicated Machine Learning Engineer specializing in large language models, advanced AI systems, and scalable machine learning solutions.
My research focuses on developing intelligent, responsible AI technologies that push the boundaries of computational intelligence.
I'm currently a graduate student working towards my Master of Science (M.S) degree in Computer Software Engineering.
Northeastern University (NEU)
Mumbai University (MU)
Here are some of my key skills, developed through various projects and professional experiences.
Extensive experience in various ML techniques, including deep learning, NLP, and computer vision. Projects like WanderWithGPT, TrialMatchAI, and CohortBuild showcase expertise in developing AI-driven solutions.
Advanced work with LLMs, RAG systems, and NLP tasks. Projects such as ChatWS AutoEval and CodeConvLLM demonstrate proficiency in cutting-edge NLP techniques.
Skilled in handling large datasets and building data pipelines. Projects like TrafficAdvisor and experience at Fidelity Investments highlight ability to work with big data technologies.
Proficient in cloud platforms (AWS, Azure) and MLOps practices. Work on integrating MLflow, Kubeflow, and building automated deployment systems demonstrates strong skills in this area.
Experience with various deep learning frameworks and architectures. Projects like Parallel Deep Learning for Image Captioning and work on fine-tuning models showcase expertise in this domain.
Skills in image processing, object detection, and image captioning. Projects like Nike Product Search Bot and work with CLIP embeddings demonstrate abilities in this field.
Strong programming skills, particularly in Python, with experience in building APIs, web applications, and backend systems. Projects consistently demonstrate software engineering best practices.
Proficiency in working with various databases, including SQL (MySQL, PostgreSQL) and NoSQL (MongoDB, Elasticsearch). Experience in optimizing database interactions and query performance.
Skills in exploratory data analysis, statistical analysis, and data visualization. Projects like MeloMuse and work on creating dashboards for various applications highlight this skill.
Here you can see some of the projects I've done on my own time.
Explore practical techniques for compressing prompts in large language models, enhancing performance and efficiency.
Understand Big O Notation through the lens of machine learning, simplifying complexity analysis.
Delve into the creation of intelligent medical diagnostics using auto-prompting and chaining techniques with DSPy.
Learn how to build more efficient AI agents for solving code challenges using AutoGen frameworks.
Step-by-step guide to deploying a language model as a Streamlit app on Google Colab for easy accessibility.
A comprehensive guide on implementing Latent Diffusion models in Python to create your own DALL-E style images.
A short summary of my work experience.
As a Graduate Teaching Assistant, I'm working on reducing hallucinations in LLMs, implementing Self-Instruct for data scaling, mentoring students in Generative AI concepts, and developing educational content for LLM research.
As a Data Scientist Co-op, I developed LLM evaluation methods, enhanced entity extraction, improved 401(k) form processing, and boosted query resolution systems using advanced NLP and machine learning techniques.
As a Senior Product Engineer, I led the development of an MLOps platform, improved ML model performance, implemented continuous A/B testing, and enhanced cross-team collaboration through integration of various tools and frameworks.
As a Product Engineer, I created an Auto-Model-Deployment system, implemented monitoring for major ML frameworks, and engineered scalable RESTful services for ML model deployment.
As a Graduate Engineering Trainee, I developed APIs for database interaction, wrote test cases, researched dockerization of code IDEs, and implemented AutoML frameworks to streamline data preparation processes.
Voices of Endorsement & Professional Validation
"Devesh demonstrates exceptional technical skills and innovative thinking. His machine learning solutions are not just technically sound, but also creatively designed to solve complex problems."
"His deep understanding of MLOps and AI frameworks has been transformative. Devesh brings a rare combination of technical expertise and strategic vision to every project."
"Beyond technical skills, Devesh demonstrates a remarkable commitment to ethical AI development. His work consistently prioritizes responsible and transparent technological solutions."