Ekaansh Khosla - Master thesis - LLM powered RAG System

Recommended expert

Erlangen, Germany

Experience

Dec 2024 - May 2025

6 months

Erlangen, Germany

Developed a RAG system to automate student queries with 96% accuracy, built using FastAPI and LangChain and deployed on the university server with Docker.
Evaluated performance using RAGAS, comparing LLMs (Llama3.3, Llama3.1, GPT-4o-mini), vector embeddings, and various retrieval techniques within the RAG pipeline.
Technical Skills: Python, FastAPI, Docker, AWS, LangChain, LangSmith, NLP, HTML, CSS

Nov 2023 - May 2025

1 year 7 months

Forchheim, Germany

Used LLM to convert unstructured complaint data from PDFs into structured Excel format.
Built a CNN-based image classifier to verify checkboxes in complaints data, achieving a 0.97 F1-score.
Optimized Power BI dashboard creation processes through ETL operations for complaints data.
Technical Skills: Python, Azure, Snowflake, LangChain, PyTorch, LLMs, NLP, Power BI, SQL, Excel

Nov 2023 - Dec 2024

1 year 2 months

Nuremberg, Germany

Performed multi-label classification on electromagnetic coil dataset, attaining an F1-score of 0.94.
Investigated Stable Diffusion img2img and Dreambooth txt2img techniques for data augmentation on the electromagnetic coil dataset, achieving a 2.1% improvement in F1-score using img2img technique.
Technical Skills: Python, PyTorch, Optuna, Stable Diffusion, Deep learning, Computer Vision, HPC

Jul 2022 - Mar 2023

9 months

Erlangen, Germany

Performed data preprocessing on marketing data, including cleaning, transformation, and normalization, ensuring analytical readiness and reducing manual effort by 45%.
Used SQL and Excel to analyze invitations sent to leads for an online marketing event.
Technical Skills: Python, SQL, Selenium, Beautiful Soup, Power BI, Excel

Jun 2019 - Dec 2020

1 year 7 months

Noida, India

Led a team of four engineers in building scalable AI/ML solutions and proof of concepts.
Developed ML solutions for Samsung smartphones, improving logs debugging and hate speech classification.
Developed and deployed ML models as REST APIs using Python, Flask, and Docker.
Implemented ML models on AWS Lambda and SageMaker for real-time and batch inference.
Designed CI/CD pipelines for model deployment and integration using GitHub Actions and AWS services.
Technical Skills: Python, AWS, MLOps, REST API, Flask, Docker, CI/CD, PyTorch, NLP

May 2018 - Jul 2018

3 months

Bengaluru, India

Implemented a predictive cooling system for Samsung smartphones, forecasting device temperature up to 15 seconds with RMSE of 1.17°C, using time series modeling techniques.
Technical Skills: Python, Scikit-learn, SciPy, Matplotlib, Pandas, Time series modeling

Prog. Lang.: Python, Sql, R, Java, C++, Html, Css, Javascript, Typescript
Mlops Tools: Git, Github, Mlflow, Docker, Ci/cd, Rest Api, Fastapi, Flask, Aws, Azure, Gcp
Ai/ml/dl/genai: Pytorch, Pytorch Lightning, Tensorflow, Keras, Rag, Llm Fine-tuning, Langchain, Langgraph, Hugging Face, Computer Vision, Nlp, Stable Diffusion, Dreambooth
Data Science And Analytics: Pandas, Numpy, Scikit-learn, Xgboost, Random Forest, Plotly, Matplotlib, Feature Engineering, Exploratory Data Analysis, Hypothesis Testing, Etl, Power Bi, Excel, Qlik
Big Data: Databricks, Apache Spark, Pyspark, Data Warehousing
Databases: Postgresql, Ms Sql, Nosql, Mongodb