AI Research Intern

• Collaborating with Prof. Amitava Das on LLM Fairness Framework targeting caste and religious bias mitigation in Indian socio-cultural contexts; developing novel bias detection algorithms for constitutional AI alignment
• Curated comprehensive AI Constitution dataset with 46K prompts spanning diverse religious and caste identities; benchmarked GPT-4, LLaMA, and Indic-specific language models for bias consistency evaluation
• AMBEDKAR: A Multi-level Bias Elimination through a Decoding Approach with Knowledge Augmentation for Robust Constitutional Alignment of Language Models (Under review at TACL 2026) — arXiv:2509.02133

Software Engineer

• Architected production-grade RAG multi-agent chatbot system with Server-Sent Events, entity extraction algorithms, and real-time widgets; delivered 35% surge in user engagement and page visit metrics
• Engineered high-performance retrieval infrastructure featuring API-level SQL converters and OpenSearch integration; accelerated query response times from 1.5s to 60ms (96% performance boost) serving production traffic
• Spearheaded News Content Platform with autonomous content generation pipeline and SEO optimization; scaled platform to serve 3K+ daily active users with intelligent recommendation algorithms
• Orchestrated intelligent Apache Airflow ETL pipelines processing 100K+ financial records daily; established automated 10-minute update cycles for CPI/CFPI metrics with comprehensive monitoring

Research Intern

• Preprocessed 8000 cancer datasets for the Integrated Healthcare Platform, enabling clustering of similar datasets through dimensionality reduction techniques
• Conducted data cleaning, integration, and clustering using BioBERT embeddings and t-SNE visualization, facilitating near-instant identification of similar datasets, thereby eliminating manual review and accelerating data integration

ML Engineering Intern

• Designed enterprise-scale Retrieval-Augmented Generation system with Qdrant vector database and LlamaIndex orchestration; enhanced regulatory document retrieval accuracy by 50% while eliminating manual processing overhead
• Deployed 10K+ financial regulatory documents into vector database with hybrid search capabilities; accelerated information retrieval efficiency by 50% through semantic similarity and keyword matching fusion

Research Intern

• Collaborated with orthodontists to develop an intra-oral monitoring system, researching sensors and creating a prototype
• Integrated circuitry and programming for real-time tracking, increasing wear time reporting accuracy by 23%
• Achieved 98% precision in varied intra-oral conditions, enhancing oral health monitoring metrics

Indraprastha Institute of Information Technology, Delhi (IIIT Delhi)

CGPA: 8.14