Headline

Machine Learning Engineer | Agentic AI & Intelligent System Design | End-to-End ML Architectures | Multimodal AI | AI/ML Researcher

About

I design & build end-to-end machine learning systems with a focus on Agentic AI and scalable pipeline architectures. My work centers on designing multi-agent, multimodal AI and fine-tuning LLM & VLM systems that integrate perception, reasoning, retrieval, and execution into cohesive, production-ready workflows. I specialize in transformer-based systems, modular ML design, orchestration strategies, and architecting scalable inference pipelines that operate reliably under real-world constraints. Core areas: • Agentic & multi-agent systems • Multimodal AI • Retrieval and reasoning architectures • Production ML deployment I aim to build intelligent systems that are architecturally sound, scalable, and optimized beyond model-level accuracy.

Location

Bengaluru

AI stack

cursor

opencode

claudecode

Work experience

AI Engineer at Mondee, Mar 2026 - Present · 6 mos, Hyderabad, Telangana, India, Agentic AI design and development

Machine Learning Engineer at HyperWorks Imaging, Aug 2025 - Feb 2026 · 7 mos, Bengaluru, Karnataka, India, Building ingeneous deep learning pipelines that inherits the cutting edge ML models and frameworks.

Machine Learning Intern at HyperWorks Imaging, Jan 2025 - Aug 2025 · 8 mos, Bengaluru, Karnataka, India, Research and Development of efficient OCR models and algorithms.

AI Intern at Aavaaz, Dec 2024 - Jan 2025 · 2 mos, 440 N Wabash Ave, Chicago, IL, 1. Assist in the development, testing, and deployment of AI models for speech and language processing. 2. Collaborate with team members to analyze data and enhance system performance. 3. Gain hands-on experience in cutting-edge AI technologies while contributing to real-world applications.

AI/ML Research Intern at MCA - NIT Raipur, Sep 2024 - Jan 2025 · 5 mos, Raipur, Chhattisgarh, India, • Performed side-channel attacks on AES-128 using AI/ML, recovering encryption keys via power consumption traces. • Optimized deep learning models (CNNs, MLPs) for trace classification, surpassing traditional DPA methods. • Evaluated cryptographic countermeasures and proposed hardware security enhancements against AI-driven attacks.

Computer Vision Intern at Shoonya Technologies Ltd., Jun 2024 - Sep 2024 · 4 mos, Bengaluru, Karnataka, India, • Improved waste sorting accuracy with advanced deep learning (YOLOv8, TensorFlow, PyTorch), achieving 85% precision in recyclable material classification. • Led a team of seven to optimize model performance using Python, Azure ML, and custom dataset preparation techniques. • Deployed the solution on Azure cloud infrastructure, ensuring scalability, robust performance, and integration for widespread application.

Education

Bachelor of Technology - BTech, Computer Science - ARTIFICIAL INTELLIGENCE, chhattisgarh swami vivekanand technical university

10+2, Science (PCMB), Delhi Public School, Siliguri

Public links

github: https://github.com/Abhisekguha