Building Intelligent Systems with Generative AI & Multi-Agent Workflows
I am an AI Engineer specializing in Generative AI, Multi-Agent Systems, and Large Language Models. With expertise in building scalable AI solutions, I'm passionate about leveraging cutting-edge technologies to solve complex real-world problems and drive innovation in the AI space.
Education
Experience
Achievements
Indian Institute of Information Technology Allahabad
Prayagraj, Uttar Pradesh
Master of Technology in Information Technology (Artificial Intelligence)
August 2022 – June 2024 | GPA: 8.73/10
Techno Engineering College Banipur
North 24 Parganas, West Bengal
Bachelor of Technology in Mechanical Engineering
August 2018 – June 2022 | GPA: 9.10/10
Software Engineer - Gen AI
Concert AI, Bengaluru, Karnataka
May 2025 – Present
Built multi-agent Text-to-SQL workflows using LangGraph and LLMs, integrating context caching and prompt engineering to deliver real-time medical dashboards with 62% lower latency and 90% cost savings. Developed LLM-based systems for automatic extraction and structuring of medical prescription information.
AI Engineer
Centific, Hyderabad, Telangana
January 2024 – May 2025
Developed agentic AI-powered RAG chatbots with integrated web search for extracting insights from 500+ pages.
Built Generative AI systems for detecting multi-identity submissions across 350,000 resumes.
Collaborated with global cross-functional teams on PEFT automation, improving LLM performance.
Developed algorithms using Kalman filters and ResNeXT-50 achieving 88.3% IDF1 on AI City 2023 dataset.
🏆 Academic Excellence
Secured 1st position in AI department at IIIT Allahabad
🥇 Competitive Programming
Ranked 474 out of 1.5 lakh participants in CodeKAZE 2023
💻 Problem Solving
Solved 1500+ DSA questions across various coding platforms within 1 year
Multi-agent pipeline using LangGraph with context caching to deliver real-time dashboards with 62% lower latency and 90% cost savings.
Insurance policy analysis from 500+ pages with web search and retrieval-augmented generation for robust answers.
Generative AI pipeline for multi-identity detection across 350k resumes using transformer embeddings and clustering.
It can Track Multiple People across multiple cameras and achieved 88.3% IDF1 on AI City 2023 leveraging ResNeXt-50, Kalman filtering, and re-identification.
Improved R² from 0.64 to 0.93 by switching from LSTM to GRU for time-series prediction.
Satellite imagery analysis with CNNs for flood event detection and mapping.
amitroy2930@gmail.com
+91-8900067947