• Worked with clients including META, XAI, and MISTRAL, delivering AI-driven solutions tailored to their needs.• Developed and delivered robust Right-Hand Fine-Tuning (RHFL) and Supervised Fine-Tuning (SFT) notebooks, customized to client specifications for enhanced model performance.• Conducted comparative analysis of advanced AI models (ChatGPT-4o, Mistral 8B, Llama 3.1, and Claude Sonnet) across multiple performance metrics to inform strategic decision-making and optimize… Show more • Worked with clients including META, XAI, and MISTRAL, delivering AI-driven solutions tailored to their needs.• Developed and delivered robust Right-Hand Fine-Tuning (RHFL) and Supervised Fine-Tuning (SFT) notebooks, customized to client specifications for enhanced model performance.• Conducted comparative analysis of advanced AI models (ChatGPT-4o, Mistral 8B, Llama 3.1, and Claude Sonnet) across multiple performance metrics to inform strategic decision-making and optimize deployment.• Assisted in improving citation accuracy and enhanced web search capabilities of LLM models, ensuring more reliable, relevant, and up-to-date references for model response. Show less

• Optimized inference performance and insights extraction using advanced LLMs: Mistral, OpenAI GPT-4, Groq. Developed a RAG pipeline for Gnews articles. Integrated technologies: Langsmith, Llama Index, Hugging Face, Knowledge Graphs. • Designed and implemented an LLM parser, reducing development time by a factor of 10.• Developed a Standardiser API wrapper, improving data quality and increasing client satisfaction by 45%.• Created Conversational AI over Elasticsearch data… Show more • Optimized inference performance and insights extraction using advanced LLMs: Mistral, OpenAI GPT-4, Groq. Developed a RAG pipeline for Gnews articles. Integrated technologies: Langsmith, Llama Index, Hugging Face, Knowledge Graphs. • Designed and implemented an LLM parser, reducing development time by a factor of 10.• Developed a Standardiser API wrapper, improving data quality and increasing client satisfaction by 45%.• Created Conversational AI over Elasticsearch data, designing LLM agents for search, aggregation, and function calling.• Automated preprocessing and formatting of data from government sources, handling 1-2 million records monthly.• Reported to the CEO, led a team of 5-6 data specialists, maintaining the legacy codebase and ensuring high standards for new code and data quality.• Established an advanced CI/CD pipeline, reducing build failures by 30% and accelerating release cycles by 25%. Engaged in pull requests and code assessments. Show less

• Orchestrated a customer assistance chatbot application utilizing AWS Lex, tailored to meet the unique needs of clients.• Implemented integration of advanced Language Models (LLMs) including GPT-3.5, LLama2 to extract domain-specific keywords from scraped text. Utilized Retrieval-Augmented Generation (RAG) techniques to evaluate and score the extracted keywords against those generated by the LLM. Tech stack included OpenAI, Hugging Face, LangChain, Chroma DB, and FAISS for vector… Show more • Orchestrated a customer assistance chatbot application utilizing AWS Lex, tailored to meet the unique needs of clients.• Implemented integration of advanced Language Models (LLMs) including GPT-3.5, LLama2 to extract domain-specific keywords from scraped text. Utilized Retrieval-Augmented Generation (RAG) techniques to evaluate and score the extracted keywords against those generated by the LLM. Tech stack included OpenAI, Hugging Face, LangChain, Chroma DB, and FAISS for vector functionalities. • Contributed to text evaluation using the RAG pipeline, increasing processing efficiency by 65% while collaborating with subject matter experts for LLM evaluation. • Engineered a Product recommendation system powered by BERT, enabling seamless user exploration across diverse categories. • Utilized TF/IDF for producing search results in our company's search engine, achieving a 90% accuracy improvement over the previous search system and reducing search result times to 2 seconds. • Automated and fine-tuned a scheduled pipeline for data cleanup, addressing issues related to invalid links and NSFW images, thereby enhancing data integrity.• Contributed to the hiring process by conducting technical interviews and providing valuable insights into candidate evaluation. Show less

• Streamlined the process of scraping from Shopify and WordPress sources through comprehensive automation. • Conducted a Proof of Concept (PoC) on sentiment analysis for product reviews across various platforms, exploring insights for enhanced decision-making.

• Conducted Proof of Concept testing on diet recommendation algorithms, assessing the efficacy of various methods including KNN, Random Forest, and other advanced ML techniques. • Reported directly to CTO: Data Engineering lead for designing and developing the web tagging portal that provides a rich visual summary of meals, plans and ingredients. Used by the data entry interns in the creation of plans and metadata. • Integrated multiple third-party APIs creating pipelines to… Show more • Conducted Proof of Concept testing on diet recommendation algorithms, assessing the efficacy of various methods including KNN, Random Forest, and other advanced ML techniques. • Reported directly to CTO: Data Engineering lead for designing and developing the web tagging portal that provides a rich visual summary of meals, plans and ingredients. Used by the data entry interns in the creation of plans and metadata. • Integrated multiple third-party APIs creating pipelines to provide on demand nutrients and food details, implementing various ETL pipelines. Show less

• Performed Data Mining and utilized various NLP techniques, including sentiment analysis and text classification, to extract valuable insights from textual data.• Built Python bots to automate data scraping processes, improving efficiency and accuracy in data collection.• Acquired proficiency in Business Intelligence (BI) tools such as PowerBI, Tableau, Kibana, and Google Data Studio.• Developed visually appealing and interactive dashboards and reports using BI tools for data… Show more • Performed Data Mining and utilized various NLP techniques, including sentiment analysis and text classification, to extract valuable insights from textual data.• Built Python bots to automate data scraping processes, improving efficiency and accuracy in data collection.• Acquired proficiency in Business Intelligence (BI) tools such as PowerBI, Tableau, Kibana, and Google Data Studio.• Developed visually appealing and interactive dashboards and reports using BI tools for data visualization and analysis. Show less