
Rohit Gupta
Verified Expert in Engineering
AI Engineer and Developer
Bengaluru, Karnataka, India
Toptal member since June 21, 2023
Rohit worked as a senior ML engineer at Prem AI, an ML back-end engineer at Luma AI, and an ML research engineer at Lightning AI, focusing on their PyTorch Lightning project. Later, he became an ML lead at Mazaal AI to work on ML pipelines and worked at Shopadvisor AI as the founding AI engineer. As a freelancer, he worked with clients on LLM-related projects and built full-stack ML projects. His tech stack includes PyTorch, Python, Git, Docker, AWS, and GCP, with ML, NLP, and speech expertise.
Portfolio
Experience
- Natural Language Processing (NLP) - 3 years
- FastAPI - 3 years
- PyTorch - 3 years
- Deep Learning - 3 years
- Python 3 - 3 years
- Machine Learning - 3 years
- Amazon Web Services (AWS) - 2 years
- LangChain - 1 year
Availability
Preferred Environment
GitHub, PyTorch, Python 3, Amazon Web Services (AWS), Natural Language Processing (NLP), LangChain, Machine Learning, Deep Learning
The most amazing...
...project I've developed is a PyTorch Lightning open-source project with over 27,000 starts on GitHub.
Work Experience
Machine Learning Engineer
Freelance
- Implemented the whole AI part of a platform to create chatbots using your documents or website with multilingual functionality (https://rafiq.ai).
- Implemented the whole AI and back end of a chatbot for US townships (https://munichat.netlify.app).
- Did some consulting jobs for companies implementing projects in the LLM space.
ML Back-end Engineer
Luma AI
- Built the back end for their upcoming projects for creative partners using FastAPI, MongoDB, and Redis.
- Integrated their in-house diffusion models on the back end.
- Created integration of custom LangChain agents on the back end.
Senior ML Engineer
Prem Labs
- Led the research department for Prem AI, which involves model training and finetuning.
- Built the LLM training and evaluation pipelines for multi-node GPU training.
- Trained their Prem-1B-chat LLM model from scratch. This includes pre-training, SFT, and model alignment using direct preference optimization (DPO).
- Researched small language models for retrieval-augmented generations (RAGs).
Founding AI Engineer
Shopadvisor AI
- Built an AI shopping advisor for eCommerce companies using LLMs.
- Led everything in product and engineering and a team of three.
- Built the whole back end and OpenAI LLMs integration.
Machine Learning Lead
Mazaal AI
- Built a no-code ML platform for users to construct, train, and deploy their ML models. Created all kinds of pipelines in the image and text domain and connected these services with AWS and Runpod for deployment.
- Deployed ML models on the serverless Runpod platform as inference APIs.
- Deployed zero-shot models for labeling tasks such as image object detection and segmentation.
- Added few-shot pipelines for text-related tasks to enable training models with a small amount of data.
Research Engineer
Lightning AI
- Deployed a text-to-image generation model in production. Implemented a load balancer with dynamic batching to improve the model serving performance from 10 to 500 concurrent users. The app hit around 8,000 requests in two days without any failures.
- Led the stable diffusion research with two team members, exploring the limits of the Lightning framework when a foundational model is deployed in production.
- Collaborated with the Colossal AI team to integrate the Colossal AI engine that implements different parallelism algorithms that are especially interesting for developing SOTA transformer models.
- Contributed to improving Tuner callbacks, fully-shared data parallel (FSDP) auto-wrappers, and DeepSpeed integration.
- Worked on and maintained a Lightning open-source project built on top of PyTorch with currently 20,000 stars. When I joined it, I was already in the top 10 as a core contributor, and now I am in the top five.
Data Scientist
EpiSource
- Reduced the processing time from days to an hour by building a complete automated pipeline to generate patient profiles for Health Reimbursement Arrangements (HRAs) and model deployment on AWS and GitHub CI/CD pipeline.
- Worked on the in-house development of the data lake warehouse using PySpark, ensuring data integrity and correctness in various verticals of business.
- Analyzed the HRA and telehealth services, thus reducing the cost incurred and optimizing operations and deployment of design and procedures for weekly reports. The analysis assisted in making major business decisions to improve the processes.
Experience
PyTorch Lightning
https://github.com/Lightning-ai/lightningEarwise
Rafiq AI
Munichat
https://munichat.netlify.app/t/moorestownEducation
Bachelor's Degree in Computer Science
Maharaja Agrasen Institute of Technology (MAIT) - Delhi, India
Skills
Libraries/APIs
PyTorch, Natural Language Toolkit (NLTK), SpaCy, REST APIs, Pandas, Matplotlib, PySpark
Tools
GitHub, Git, ChatGPT
Languages
Python 3, Python, SQL
Paradigms
Automation, Distributed Computing
Platforms
RunPod, Docker, Amazon Web Services (AWS)
Storage
Amazon S3 (AWS S3), MySQL, Redis, PostgreSQL, MongoDB
Frameworks
Streamlit
Other
Natural Language Processing (NLP), LangChain, Machine Learning, Deep Learning, FastAPI, Artificial Intelligence (AI), Language Models, OpenAI GPT-4 API, Generative Pre-trained Transformers (GPT), Generative Pre-trained Transformer 3 (GPT-3), Data Science, Chatbots, Minimum Viable Product (MVP), Data Scraping, Text Classification, OpenAI GPT-3 API, APIs, Hugging Face, Frameworks, API Integration, OpenAI, Security, Pinecone, Neural Networks, Data Scientist, Algorithms, Chatbot Conversation Design, Team Leadership, Conversational Interfaces, IT Automation, Software Development, Speaker Identification (SI), Coherent UI, Weaviate, Large Language Models (LLMs), Open-source LLMs, Diffusion Models, Stable Diffusion, Diffusion-based AI Models
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring