Headline

AI/ML Engineer @ HARMAN International | GenAI • Agentic AI • Multimodal LLMs • RAG • AWS • Deep Learning • NLP • Computer Vision

About

I’m an AI/ML Engineer currently working at HARMAN International, where I build and deploy AI systems that actually get used in production. Most of my work revolves around LLMs, multimodal models, and automation. I’ve worked on things like multi-agent systems, video understanding pipelines, and AI-driven workflows that reduce manual effort and make processes faster and more reliable. I enjoy working on problems where AI meets real-world systems—whether that’s designing end-to-end pipelines, improving model performance, or making solutions scalable and practical. Over time, I’ve developed a strong interest in: • LLMs and RAG-based systems • Multimodal AI (vision, text, audio) • Building production-ready ML systems • Exploring how agents can automate complex tasks I’m currently exploring opportunities in AI/ML where I can continue building impactful systems and learn from strong teams. Always open to connecting, discussing ideas, or collaborating on interesting problems.

Location

Bengaluru

AI stack

claude code

claude

elevenlabs

hugging face

langchain

langgraph

Work experience

Associate AIML Engineer (R&D) at HARMAN International, Jul 2025 - Aug 2026 · 1 yr 2 mos, Bengaluru, Karnataka, India · Hybrid, • Working on building production-grade AI systems focused on multimodal intelligence and agentic workflows • Developed a multi-agent video redaction platform with AI-driven PII detection, reducing manual effort by ~70% • Designed AI-powered document intelligence pipelines to automate accreditation workflows and reduce manual intervention by ~60% • Migrated legacy systems to a Python-based AI backend, improving performance and reducing technical debt • Built industrial AI solutions for anomaly detection and predictive maintenance, reducing downtime by ~30%

AI Engineer Intern (R&D) at HARMAN International, Nov 2024 - Jun 2025 · 8 mos, Bengaluru, Karnataka, India · Hybrid, • Worked on multimodal AI for circuit design (EDA) using LLaVA & LLaMA models • Built an AI assistant for engineers, improving productivity and design efficiency • Applied LoRA and cross-attention techniques for efficient model fine-tuning • Contributed to a video-language segmentation system (RVOS) with strong real-world accuracy • Explored multimodal reasoning across video and text for intelligent automation use cases

Software Engineer Intern at Senseoriginal Technologies Pvt Ltd, Jul 2023 - Jun 2024 · 1 yr, Gurugram, Haryana, India, • Developed scalable Android applications using Java, XML, Kotlin, and Jetpack Compose, enhancing user interface and logic. • Built iOS applications with Swift and SwiftUI, prioritizing responsive design for optimal user experience. • Created an NFC-based product authenticity system, ensuring secure chip scanning and verification across platforms.

KWOC'22 Contributor at Kharagpur Open Source Society, Dec 2022 - Jan 2023 · 2 mos, India

Education

Bachelor of Technology - BTech, Computer Science(artificial intelligence and machine learning), PSIT Kanpur (Pranveer Singh Institute of Technology)

Public links

github: https://github.com/ashupandey1620