Open to new opportunities in the field of NLP/CV and Machine Learning.
I am glad that you have visited my blog. I hope you find it interesting and informative!
Welcome to my page! I am Siddhesh Dosi, a post graduate from IIT Gandhinagar in Computer Science and Engineering under the mentorship of Prof. Dr. Mayank Singh.
Current Research projects:
Working on the completion task of 3D partial point cloud in vision domain.
- Experimenting with various PCN and GAN based architectures for completion problem in 3D point cloud.
Explore More
Current work-profile:
Working as full time ML/AI lead engineer at Neuroreef Labs
- Developing and deploying user specific AI-Copilots for various medical residents in USA.
Explore More
Timeline
I have joined chapter apps as a full-time Gen AI SDE engineer.
It is a merger of Got-It-AI and ChapterVitamins. I will be working on the development of the AI-based chatbots for different clients including but not limited to (Aditya Birla Capitals)[https://www.adityabirlacapital.com], JCB India, etc. There are a lot of improvements to be done in the current architecture and I am excited to work on them.
I have joined Neuroreef Labs as a full-time ML/AI Team Lead of a team of 5.
I am excited to work with a team of talented individuals and contribute to the development of cutting-edge AI solutions.
We provide custom user specific Medical Template for final notes like (Neurology, Cardiology, etc).
I have successfully completed my M.Tech in Computer Science and Engineering from IIT Gandhinagar with a CGPA of 9.33/10.
My major acdamics courses are:
- Machine Learning
- Deep Learning
- Natural Language Processing
- Algorithms
- Artificial Intelligence
- Computational Neuroscience
I have successfully defended my M.Tech thesis on “Optimizing Smaller Variant of LLMs” under the guidance of Dr. Mayank Singh.
Trained a mistral architecture based mini llm of 1B parameters on a web scrapped and processed Hindi dataset, Ganga-1B
Curated and processed the scientific papers from arXiv to train a mini scientific LLM.
Reduced the bibliographic biasness, as in Galactica, by deduplication and filtering of the dataset.
Developed and designed the core architecture of two main products
> MedAura.ai : An evidence-based medical chat bot with minimal document search time with unqiue table creation approach.
> Carecortex.ai : A full-fledged AI - copilot for doctors to assist them in their daily routine.
| - Conversation Transcription
| - Evidence-based insights generation
| - Final notes generation
| - Medical coding (ICD-10, CPT, etc)
Used aws lambda to invoke apis triggered on s3 bucket events for real-time processing of the data, independent of the browser live session.
From back-2-back interviews in smartsense, citi and broadridge, I have been selected in citi as a Software Developer Analyst.
Got a great experience in the interview process and learned a lot from the interviewers.
Though the toughest interview was of smartsense, I found citi had very thoughful and interesting interview process.
Participated in the India-level hackathon at ThirdAI Corp and won the first prize among 70+ teams.
Developed a in-depth contexual based document search engine for Google Drive, TEGD (ThirdAI Engine for Google Drive) using NeuralDB architecture.
Joined ThirdAI Corp as a Machine Learning/AI Intern.
Trained many ThirdAI’s BOLT based models in distributed environment from ParamAnanta supercomputer on
- Stackoverflow dataset
- Amazom 3M review dataset
- Arxiv abstracts dataset
- Food UDT dataset
Delivered a high performance model on Amazon-3M with Metrics (P@1=0.695, R@5=0.797, R@10=0.814) in its category
Started my journey as a M.Tech student in Computer Science and Engineering at IIT Gandhinagar. I am excited to learn and explore the field of AI and Machine Learning.
Got offers (in the first round only) from
- IIT Gandhinagar (accepted)
- IIT Jodhpur
- IIT Indore
- IIT Gawhati (spot round)
- IIIT Delhi
- IIIT Bangalore
Based on my GATE score, I have been shortlisted for the interview round in IB as Assistant Central Intelligence Officer Grade-II/Executive. I went through the final round of the interview but unfortunately couldn’t make it to the final list.
I have scored 640 in GATE-22 (AIR 1012) in Computer Science and Engineering.
I am happy with my performance and looking forward to applying for M.Tech programs.
This was my second attempt and am happy with my performance and looking forward to applying for M.Tech programs.
Graduated with a B.Tech in Information Technology with 80%.
This marked the end of an exciting chapter filled with learning, growth, and exploration of the tech world.
Secured a position as a Software Developer at InTimeTec.
It was an exciting moment to transition from a student to a professional in the software industry.
I learnt a lot of concepts related to GO-lang, go-routines and Transact-SQL/SQL Server.
Achieved a GATE score of 464 in Computer Science and Engineering, securing an All India Rank of 6028.
This was my first attempt at GATE, which set the foundation for my future academic journey.
Became a semifinalist in the prestigious e-Yantra Robotics Competition organized by IIT Bombay. Worked on solving real-world problems using robotics and embedded systems, gaining valuable hands-on experience.
Designed and implemented a working prototype of supply - line follower robot using Arduino and sensors.
Used Computer vision to process the live imaging of the bot for the navigation and real-time decision making.
Embarked on a journey into the world of technology by enrolling in B.Tech in Information Technology at SKIT & MG, Jaipur.
It was the beginning of my passion for computer science and engineering.
Deep Learning
XLSTM
Point Cloud
Point Cloud Network
Data Science
NLP
Data Curation
HPC server
Distributed training
tokenization
Pytorch
Mistral
LLM pretraining
Scientific domain data
fine-tuning
Huggingface
BBL deduplication
Python
Decision Tree [Classifier and Regressor]
Linear Regreesion
Ensemble Learning
Gradient descent
No matching items
Back to top
