Jun 2024 – Jun 2025 · 1 yr · Bengaluru, Karnataka, India · On-site

• ● Conducted research in time series forecasting and image classification, delivering advanced solutions for predictive analytics and visual recognition.
• ● Specialized in Generative AI Diffusion Models for text-to-image and image-to-image tasks, leveraging LoRA (Low-Rank Adaptation) to enhance performance on custom datasets.
• ● Built an automated labeling system for fashion products using OpenAI GPT-4o API, later optimizing with batch processing to cut costs by 60% while maintaining accuracy.
• ● Fine-tuned Stable Diffusion models with Kohya SD scripts and GUI, improving text-prompt-based image generation for fashion design.
• ● Used SD WebUI/Automatic1111 for model text-to-image inference.
• ● Designed and developed dynamic APIs using Flask to enable seamless data transfer to the Trendgully frontend, integrating dynamic filtering logic for enhanced real-time data visualization and user interactions.
• ● Pioneered the use of local system AI annotation/labelling for fashion products, leveraging open-source models such as LLaMA 3.2 Vision Instruct and Pixtral for scalable, cost-efficient annotation processes (ongoing research).
• ● Developed an AI chatbot for fashion product queries, including a summarization feature to simplify detailed analytics, saving user time.
• ● Demonstrated expertise in Generative AI, Stable Diffusion, OpenAI tools, and computer vision for fashion-tech innovations.

Aug 2023 – Mar 2024 · 7 mos · Bengaluru, Karnataka, India · Remote

• ● Built and refined innovative time series algorithms for one of the largest fashion brands to identify and segment 6 fashion trends, providing customers and retailers with actionable insights.
• ● Implemented advanced clustering and decomposition techniques to categorize time series data into 6 meaningful clusters, enabling retailers to gain granular insights into product trends and seasonality.
• ● Leveraged time series seasonal decomposition to classify time series data into 6 predefined Google Trends behaviors, providing retailers with predictive insights into the growth or decline of fashion trends.
• ● Spearheading a solution design for the jewellery brands demand prediction project.
• Skills: Time Series Decomposition, Clustering, Time Series Clustering, Time Series Forecasting, Artificial Intelligence, Data Science, EDA

Jan 2023 – Jul 2023 · 6 mos · Delhi, India · Remote

• ● Developing various machine learning and NLP (Encoder-Decoder) models to help advertisers make better decisions and get better returns.
• ● Initialized Keras Functional Api to add additional embedding data to the encoder decoder model for better predictions and to help advertisers get better returns.
• ● Continued data extraction by mongodb to fetch the required data from adbrew database for model building and data analysis.
• ● Worked on the time series analysis techniques through LSTM Encoder Decoder for the sales and spend predictions.
• ● Started with various AI/ML ,deep learning techniques like multi layer perceptrons , RNN, GRU, Single LSTM, Multi LSTM and also used FB Prophet, Neural Prophet libraries by Facebook to predict the sales and spend, but finally ,found the best model as NLP Encoder Decoder with LSTM cells.Optimizing the code using PySpark.
• ● Worked on agile methodologies software such as Kanban board, attending scrum meeting, etc.
• Skills: Keras Functional API · Statistics · Exploratory Data Analysis · Start-ups · Research · Machine Learning Algorithms · Computer Science · Amazon Web Services (AWS) · PyTorch · Analytical Skills · Time Series Analysis · Data Science · Python (Programming Language) · Data Analysis

May 2022 – Aug 2022 · 3 mos · Remote

• ● Worked on the TVO travel agency project.
• ● Directed a project where we have to analyze the patterns of the travel agency websites.
• ● Building new parameters for a travel agency website for more choice options to the users.
• ● Researched about different AWS hosting servers and extracted the information which server will provide the best efficiency for a particular project. We dealt with the issues of the active users per day, per month.

Apr 2022 – Apr 2022 · 0 mo · Remote

• ● Administered the NBA dataset for pointing the unique insights from the dataset that would be beneficial for the players in the upcoming matches.
• ● Researched and arranged the past trends of the NBA players and saw which player can perform better in the upcoming matches.
• ● For 3d visualization ,used PLOTLY.

Feb 2022 – Feb 2022 · 0 mo · Remote

• ● Handled on the IPL dataset. Analyzed the IPL dataset and got the unique insights from the dataset that would be beneficial for the players in the upcoming matches.

Sep 2021 – Nov 2021 · 2 mos · Remote

• ● Created the machine learning project through which any student can predict his/her marks in a particular exam/subject.
• ● Developed a Machine Learning project using which we can classify the fake news from a collection of news data.
• ● Accomplished completing projects using Machine Learning algorithms which we can classify the spam ham mail from a collection of mails text data.