Yashvardhan Patil

CEO

Karad, Maharashtra, India2 yrs 4 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in Machine Learning and Deep Learning techniques.
Proven experience in fine-tuning advanced AI models.
Strong background in teaching and guiding students.

Stackforce AI infers this person is a Deep Learning and AI specialist with a focus on educational and research applications.

Contact

Skills

Core Skills

StrategyMachine LearningTeachingDeep LearningText To Image GenerationImage ProcessingFine TuningContent GenerationNatural Language Processing

Other Skills

Agile MethodologiesAlgorithmsAnalytical SkillsAnalyticsApplied ResearchArtificial Intelligence (AI)Artificial Neural NetworksAttention to DetailBERT (Language Model)Business AnalysisBusiness RequirementsBusiness ReviewsBusiness StrategyC (Programming Language)C++

Experience

2 yrs 4 mos

Total Experience

1 yr 2 mos

Average Tenure

2 yrs

Current Experience

Ola electric

Assistant Manager, CEO's Office

Jun 2024 – Present · 2 yrs · Bengaluru, Karnataka, India · On-site

Strategy and Operations in the Electric divison
ML Researcher at Krutrim AI: Training ASR models for Indian Languages

StrategyOperationsMachine Learning

Indian institute of technology, delhi

Teaching Assistant

Jan 2024 – May 2024 · 4 mos · Delhi, India

Guiding 120+ students in the course ELL205 Signals and Systems

TeachingGuidance

Sonyai

Deep Learning Intern

May 2023 – Jul 2023 · 2 mos · Tokyo, Japan · On-site

Worked on Text to 2.5-D Image generation with the help of Diffusion models.
Modifed the architecture of Difusion based DeepFloyd IF (Text to 2D Image) base model to jointly generate 2.5-
dimensional images (2D Image + corresponding depth map) from text prompts to help in 3D rendering of the images.
Enabled all the 3 stage cascaded models to give 6 channel images by modifying the codebase. Worked on weight
initialization for new layers while maintaining the weights of original layers. Structured the dataset in required format.
Succesfully Employed Dreambooth and Lora fnetuning by altering codebase to make it compatible for multiple prompts

Deep LearningText to Image GenerationDiffusion Models

Shorthills ai

Deep Learning Intern

Jun 2022 – Jul 2022 · 1 mo · Gurugram, Haryana, India · Hybrid

Worked on Fine Tuning GPT Neo for content generation of Buying Guides for various commercial products.
Segregated various domains of data and formatted the training data to make it trainable on GPT-Neo's 2.7 Billion model.
Used the Hugging face's Happy Transformer and DeepSpeed library to train the GPT-Neo 2.7B model on training data.
Finetuned GPT-NEO's 2.7 Billion parameters model by altering default training hyperparameters by avoiding certain
unwanted patterns to generate buying guides for given prompts on the RTX3090 machine using remote server access.

Fine TuningGPT-NeoContent Generation