Aarush Dev

Results-driven AI/ML Engineer and Data Scientist with hands-on experience designing production-grade deep learning systems, agentic RAG architectures, and multi-agent AI pipelines. Combines rigorous ML theory with pragmatic engineering to deliver measurable business impact. Adept at reading cutting-edge research and rapidly translating it into novel solutions for enterprise bottlenecks.

Brings a multidisciplinary perspective shaped by competitive quizzing, classical arts, philosophy, and martial discipline, fuelling human-centred, first-principles thinking in every technical challenge.

AI Backend Engineer Intern at Pinnacle Future Build (2025-05 – Present)

• Collaborated within the R&D team migrating key enterprise services from a legacy PHP/Laravel implementation to a modern Python-based system design, ensuring zero disruption to the broader codebase.
• Designed and built FastAPI endpoints implementing clean controllers and business logic to orchestrate services between client, application, and database tiers, enabling front-end integration without router or front-end changes; delivered thoroughly tested (Postman), plug-and-play, backward-compatible APIs for incremental migration.
• Identified and resolved critical technical bottlenecks in MCP and RAG systems by reading and applying cutting-edge research, designing novel solutions that improved system performance and reliability.
• Worked on computer vision applications; helped develop vision models for training on BIM drawings, and built a novel image augmentation pipeline combining Albumentations, LoRA fine-tuning, and Real-ESRGAN.
• Gained deep experience in enterprise system design, control flow navigation, data flow modelling, and logic flow architecture across large-scale production codebases.

Research Intern at Jadavpur University (2025-04 – 2025-07)

• Developed a novel ML-enhanced correction factor methodology bridging the 0D–3D thermal model accuracy gap for battery thermal safety prediction.
• Achieved a 99.9% accuracy improvement in thermal model fidelity, translating to 91% real-world accuracy for thermal runaway prediction in real-time battery safety systems.

BS in Data Science and Applications – Indian Institute of Technology Madras (2024-05)

B.Tech in Computer Science and Engineering – Kalinga Institute of Industrial Technology (KIIT) (2023-07)

Higher Secondary Education – Delhi Public School Ruby Park (2008-04 – 2023-03)