Über Varad
A curious, product-minded engineer who enjoys transforming imaginative ideas into practical technical solutions that deliver real user value. With a foundation spanning software engineering, data science, machine learning, NLP, recommendation systems, and generative AI, and a strong learn–adapt–build mindset. Comfortable working in cross-functional teams and exploring open-ended problems while taking ownership of impactful technical solutions.
Deutsch
Konversationssicher
Englisch
Verhandlungssicher
Hindi
Verhandlungssicher
Marathi
Muttersprachlich oder zweisprachig
Nur remote
Führt Projekte hauptsächlich remote aus
Projekt- und Berufserfahrung
- UnternehmerTUM GmbHAIEngineerSeptember 2025 - November 2025 (2 Monate)Built an AI-driven WebXR/VR product for LSBG Hamburg to visualize future bus stop designs using BIM and LiDAR point-clouddata, supporting urban infrastructure planning decisions.Developed spatial data workflows for point-cloud processing, BIM interpretation, and interactive 3D visualization using WebXRand Three.js in virtual reality environment.Implemented a CI/CD pipeline (GitHub Actions + Coolify) enabling automated testing and continuous deployment of theproduct prototype.Conducted user interviews, usability testing, and stakeholder interviews with city planning teams to identify engineering painpoints and iteratively validate product direction.Communicated complex AI and spatial computing concepts to non-technical stakeholders, aligning technical feasibility withplanning and design requirements.Technologies: WebXR, Three.js, FastAPI, LiDAR / Point Cloud Data, Blender, CloudCompare, GitHub Actions, Coolify, Git, Jira,Figma
- University of Hamburg,ResearchAssistant-NLPSeptember 2024 - Februar 2025 (5 Monate)Contributed to research on open-source automated annotationof event concepts in literary texts, supporting the analysis ofnarrative structures across large text corpora.Analyzed genre-specific event patterns across multiple narrativedatasets using statistical methods to uncover structuraldifferences in storytelling.Developed an interactive visualization tool (React, ApacheECharts) to map character co-occurrence networks in novels,enabling exploratory analysis of narrative relationships.Trained a T5 transformer model for text pseudonymization andEnglish-German translation, supporting experiments inautomated text processing and privacy-preserving NLPworkflows.Technologies: Python, NLP, T5 Transformers, React, ApacheECharts, Statistical Analysis
- Cognizant Technology Solutions,ProgrammerAnalystFebruar 2022 - September 2023 (1 Jahr und 7 Monate)IndiaDeveloped and deployed a full-stack application usingJava, React, Spring Framework, AWS, and SQL,contributing to the Cognizant Development Engineer(CDE) program.Collaborated on a team project to implement robustauthentication and authorization features,strengthening application security and usermanagement
Empfehlungen
Sei die erste Person, die Varad empfiehlt
Teile Deine Erfahrung aus der Zusammenarbeit mit diesem Freelancer.
Diese Freelancer passen auch zu Ihren Kriterien
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Ausbildung und Abschlüsse
- B.Tech Information TechnologyDr.BabasahebAmbedkarTechnologicalUniversity,India2022
- M.Sc. Data ScienceTU Hamburg2025Grade: 2.3 Course Work : 1. Advance Machine Learning 2. Artificial Intelligence 3. Big Data 4. Statistical Models 5. Probabilistic Machine Learning 6. Deep Learning for Social Analytics 7. Cyber Security for Data Science 8. Secure Software Engineering 9. Machine Learning for Physical Systems 10. Intelligent System in Medicine Thesis : Multi-Objective Music Recommendation System Objective: To redesign music recommendation systems by optimizing exposure for emerging artists while maintaining listener satisfaction. Methodology: Developed a multi‐objective optimization framework combining content‐based and collaborative filtering (matrix factorization) approaches, leveraging the Million Song Dataset for feature extraction and refined via Bayesian Optimization (Optuna). Results: Achieved 5× improvement in genre precision, 1.9× increase in emerging‐artist hit rate, and near‐perfect proportional exposure (EAEI=1.02), all while preserving recommendation quality and listener engagement.