Headline

AI Engineer

About

Aspiring deep learning enthusiast passionate about AI. Actively learning and implementing research papers related to deep learning. Looking to collaborate with professionals and professors in the field to pursue research opportunities and contribute to the advancement of AI.

Location

Greater Hyderabad Area

AI stack

codex

cursor

Work experience

AI Engineer at Amrita Center for Cybersecurity Systems & Networks, Aug 2024 - Present · 2 yrs 1 mo, Kerala, India, • Built an AI-powered assessment platform enabling teachers to generate exam questions 3×faster from PDFs, videos, PPTs, and YouTube content using a multi-stage RAG pipeline. • Defined multi-level question difficulty by creating custom evaluation datasets and tuning retrieval, prompting, and validation workflows, reducing LLM inference cost by 50%. • Developed scalable backend systems with Django REST Framework and FastAPI, Celery and RabbitMQ for background task processing, relational databases, S3 storage, Docker, CI/CD, and Coolify; increased test coverage from 28% to 81% supporting 120+ active users with improved reliability through memory and worker optimizations. • Architected a NIRF research intelligence platform enabling conversational querying of institutional research documents, comparative analysis, visualizations, and source-backed answers through multi-agent orchestration. • Built a custom multi-agent orchestration system with Router, Search, Library, Analytics, and Synthesis agents using MCP (Model Context Protocol)-based FastAPI microservices. • Built a pipeline that converted NIRF PDFs into structured data and artifacts, used BM25 search and PostgreSQL-based analysis to answer research queries, and returned clear, source-backed responses with citations and visualizations.

AI Engineer Intern at Aveta.ai, Aug 2023 - Jan 2024 · 6 mos, Palo Alto, California, United States, • Build a medical ICD coding system using RAG, Pinecone vector databases, attention manipulation, and LLM fine-tuning, achieving 93%+ accuracy. • Performed LoRA fine-tuning of BioGPT and MedAlpaca using HuggingFace/PEFT to adapt models to medical terminology and hospital record formats. • Processed large-scale unstructured medical PDFs (patient records, discharge summaries, billing documents) for downstream LLM workflows.

Member at amFOSS, Aug 2022 - Nov 2025 · 3 yrs 4 mos, Kollam, Kerala, India

Research Intern at ACM Student Chapter , Amrita Amritapuri, Jul 2022 - Oct 2022 · 4 mos, Kollam, Kerala, India

Education

Bachelor of Technology - BTech, Artificial Intelligence, Amrita Vishwa Vidyapeetham

Intermediate Education, FIITJEE

Public links

github: https://github.com/vishalj0501