CV

Universidad Autónoma de Madrid

Pázmány Péter Katolikus Egyetem

Grade: 4.8/5

• Basic Image Processing Algorithms (Core)
• Biomedical Signal Processing (Elective)
• Data Mining and Machine Learning (Core)
• Fundamentals and Basic Tools for Deep Learning (Core)
• Design Patterns (Elective)
• Multimodal Sensor Fusion and Navigation (Core)

Atma Ram Sanatan Dharma College, Delhi University

• Led development of Retrieval-Augmented Generation (RAG) pipelines for scientific Q&A and spectroscopy queries, integrating document ingestion, embedding generation, and Milvus vector database with automated indexing from cloud storage (S3 & Google Drive).
• Engineered scalable async APIs and NL2SQL pipelines that converted natural language queries into SQL, enabling seamless metadata/file retrieval, automated database indexing, and secure cloud file access.

• Led development of multimodal AI solutions for detecting hateful content on social media achieving SOTA results on existing benchmarks.
• Extended methodology to multi-state cases using a contrastive learning approach.

• Contributed to 2 GenAI + RAG projects for confidential industry clients.

• Worked on prototyping a chatbot for solving user queries and product recommendations using Botpress for UrbanLyfe App. Experimented with fine-tuning LLM models for enhancing response.
• Explored and brainstormed the utilization of ML methods and BI dashboards for the logistics tracking and cost optimization products.

• Worked on Speech Processing Models.

• Balanced code-mixed text data for unbiased sarcasm classification using four text augmentation and three data balancing techniques.
• Fine-tuned EfficientNetB7 on Alzheimer's MRI scans with transfer learning, achieving improved accuracy for early diagnosis.
• Developed a text detection and localization system using Transformers.

• Evaluated scalability, throughput, and latency of Solidity-based blockchain systems using benchmarking and profiling techniques.
• Currently writing a book on fundamental principles of blockchain technology.
• Researched applicability of Blockchain in Intrusion Detection Systems for CPS, UAVs and IoT systems.

• Researched 'enabled transfer learning for low-resource languages'. Fine-tuned BERTs for domain-specific tasks.
• Conducted statistical text analytics and data augmentation for various NLP tasks including hate speech detection.
• Contributed to spear-phishing analysis on UKM students using simulated emails mimicking real attacks.