Experience
Master's Thesis
Heinrich Heine Universität
- Title: Enhancing Syntactic Awareness in Transformer Language Models for Hindi Dependency Parsing
- Investigated syntactic knowledge captured by transformer language models (RoBERTa, XLM-RoBERTa) for Hindi dependency parsing, a morphologically rich and low-resource language.
- Developed structure-aware model variants (Struct_Roberta_hi, Struct_XLMR) by integrating a CNN-based parser network between transformer layers, inspired by the StructFormer architecture.
- Conducted extensive error analysis including label-wise, distance-based, direction-based, sentence length-based, and LVC/Non-LVC evaluations.
- Evaluated models on downstream NLP tasks (NER, POS tagging) using the IndicXTREME benchmark.
Pretraining and Evaluation of Language Models
Heinrich Heine Universität
- Pretrained language models Roberta, StructFormer from scratch on large-scale Hindi text corpus.
- Evaluated on downstream tasks such as POS tagging, NER, sentiment analysis.
- Compared performance with baseline models like IndicBERT using standard metrics (accuracy, F1-score).
Sarcasm Detection in Nepali and Hindi Text
Heinrich Heine Universität
- Fine-tuned pretrained NLP models from Hugging Face on Hindi and Nepali text for sarcasm detection.
- Performed data collection, cleaning, and preprocessing.
- Evaluated models using accuracy, precision, recall, and F1-score; visualized results with matplotlib and seaborn.
ETL Pipeline for MongoDB to PostgreSQL
- Designed and implemented an ETL pipeline for MongoDB collections.
- Extracted, transformed, and loaded JSON data into PostgreSQL using Python.
- Developed and executed SQL script in PostgreSQL.
- Created data analysis and visualization reports using pandas, networkx and matplotlib.
Movie Data Analysis
- Analyzed a movie dataset to uncover trends and insights in film attributes.
- Imported data into PostgreSQL and executed SQL queries.
- Cleaned and preprocessed data to ensure accuracy and created visualizations.
Linear Regression Model
- Developed a Linear Regression model to predict outcomes based on various input features using Python.
- Implemented data preprocessing, cleaning and visualization.
- Evaluated model performance through metrics such as R-squared and Mean Squared Error (MSE).
Image Registration
Heinrich Heine University
- Applied image registration techniques to align unregistered brain images from a research group's dataset using Python.
- Utilized feature detectors provided by scikit-image and OpenCV to detect image features.
- Successfully registered a batch of brain images by linking transformations together.
- Employed the RANSAC method to remove outliers from the matched image features.
Studio Warehouse Associate
Shop LC GmbH
- Coordinated studio inventory operations and supervised warehouse team members.
Spectral Graph Theory and Graph Signal Processing
Heinrich Heine University
- Applied Time Evolution Fourier series techniques to analyze traffic data and identify patterns using Python.
- Used mathematical models to uncover trends and fluctuations in traffic over time.
- Generated a graph from traffic sensor data using a similarity function based on sensor location distances.
IT Officer
Stopgrab E-commerce Company
- Developed and executed social media and digital marketing strategies; managed multi-platform company profiles.
- Produced graphic and video content using Adobe Photoshop, Illustrator, and After Effects.
Prasnottar
NCIT
- Developed an interactive e-learning class module mobile application using Java for the frontend and Django REST Framework for the backend.
- Created a user-friendly interface within Android Studio IDE.
- Established database in Google Firebase to handle data storage, retrieval, and manipulation.
Evento
- Created "Evento," an e-commerce platform for special events, using Java for both frontend and backend development within Android Studio IDE.
- Integrated Google Firebase to manage product information, user data, and order details.
Chitraka
NCIT
- Developed "Chitraka," a platform for photographers to showcase their work and compete, using Java for both frontend and backend operations within the Android Studio IDE.
- Integrated Google Firebase as the database management system to store and retrieve user data, including photos, user profiles, and preferences.
Industry experience
See where this freelancer has spent most of their professional time.
Experienced in Retail, Information Technology, and Education.
Business area experience
See which departments and functions this freelancer has contributed to most.
Experienced in Logistics, Information Technology, Marketing, Product Development, and Research and Development.
Frequently asked questions
Have questions? Find more information here.
Saruna is based in Duisburg, Germany.
Saruna speaks the following languages: Nepali (Native), English (Advanced), Hindi (Advanced), German (Elementary).
Saruna has at least 6 years of experience. During this time, Saruna has worked in at least 13 different roles and for 5 different companies. The average length of individual experience is 1 year and 6 months. Note that Saruna may not have shared all experience and actually has more experience.
Based on recent experience, Saruna would be well-suited for roles such as: Master's Thesis, Pretraining and Evaluation of Language Models, Sarcasm Detection in Nepali and Hindi Text.
Saruna's most recent position is Master's Thesis at Heinrich Heine Universität.
In recent years, Saruna has worked for Heinrich Heine Universität, Heinrich Heine University, and Shop LC GmbH.
Saruna is most experienced in industries like Retail, Information Technology, and Education. Saruna also has some experience in Media and Entertainment and Healthcare.
Saruna is most experienced in business areas like Logistics, Information Technology, and Marketing. Saruna also has some experience in Product Development, Research and Development, and Business Intelligence.
Saruna has recently worked in industries like Retail, Information Technology, and Education.
Saruna has recently worked in business areas like Logistics, Information Technology, and Research and Development.
Saruna holds a Master in Data Science And Artificial Intelligence from Heinrich Heine Universität and a Bachelor in Computer Engineering from Nepal College of Information Technology, Pokhara University.
Saruna is immediately available full-time for suitable projects.
Daily rate distribution
The rates shown represent the typical market range for freelancers in this position based on recent contracts on our platform.
Average rates for similar positions
Rates are based on recent contracts and do not include FRATCH margin.
The average daily rate is the mean of all daily rates from recent contracts of comparable freelancers on our platform.
The median daily rate is the middle value of all daily rates — half of comparable freelancers charge less, half charge more. Unlike the average, it is barely affected by outliers.
Calculated based on our freelancers’ daily rates as of 19 Nov 2026. Actual rates may vary depending on seniority level, experience, skill specialization, project complexity, and engagement length.
Similar freelancers
Discover other experts with similar qualifications and experience
Ekaansh K.
Last position:
Master thesis - LLM powered RAG System at Friedrich-Alexander-Universität Erlangen-Nürnberg
- 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
Noushiq M.
Last position:
Projects at Institute for Intelligent Systems
- Evaluated and analyzed camera-based traffic light and sign recognition systems across various LLM-based autonomous driving systems (LMDrive, BEVDriver)
- Implemented a VLM-based traffic notice instruction generation unit for a closed-loop autonomous driving system that alerts the driver to unexpected driving incidents
- Developed an independent LLM-based local chatbot using Llama, DeepSeek, and Qwen, including an MLflow evaluation framework
Kamal R.
Last position:
Design and development of a requirements verification tool based on the INVEST framework and artificial intelligence at Ranasoft Internet Technologies
- Design of requirement verification based on the INVEST principle
- Implemented the forecasting model using scikit-learn and BERT
- Customized the T5 model to improve user stories according to the INVEST framework
- Built the web client with ReactJS and set up communication with the backend using FastAPI
- Obtained NLP and scikit-learn certifications
Kiran Kumar K.
Last position:
Applied NLP: Word-Level Encoding for Smarter Event Predictions at University of Siegen
- Engineered 12+ Seq2Seq models (LSTMs/GRUs) to train an AI model, supporting AI Agent Evaluation Analyst and online projects in complex systems.
- Conducted 15+ experiments to improve forecasting accuracy by 28%, applying analytical thinking and testing models for QA and edge case coverage.
- Researched 20+ papers as a consultant to guide Large Language Model design, ensuring logical implications and domain of expertise alignment.
- Saved 40% compute time via model compression with reusable PyTorch framework, aiding developers and writers to suggest refinements and improve policies.
Srijan M.
Last position:
Senior Product Manager – Revenue Management at SIXT SE
- Built and scaled a data-science price engine across 8 EU markets, lifting fleet margin by ~2% on a €1.5B+ base
- Launched a generative-AI insights platform in 29 countries, cutting manual analysis by ~40% and driving weekly actions
- Led a hybrid team and aligned 100+ stakeholders to refine pricing logic and accelerate rollout across regions
- Set up KPI governance with WBR/MBR rhythms, reducing decision latency by ~30% across European pricing teams
Shyam Sundar R.
Last position:
GenAI Engineer at Freelance
- Built a hybrid semantic and keyword search and LLM-based requirement extraction from conversational queries, boosting search accuracy by 85%, cutting zero-result searches by 70%, and reducing search time by 60%.
- Deployed a production-ready API with monitoring dashboards over 100K+ products, keeping response times under 2s and reducing customer search-to-purchase time by 40%.
- Technologies: Python, BGE-M3, Qwen2.5, FastAPI, Qdrant, Meilisearch, Docker, Prometheus, vLLM.
Experts recently working on similar projects
Freelancers with hands-on experience in comparable project as a Master's Thesis
Nearby freelancers
Professionals working in or nearby Duisburg, Germany

