Jaya Soorya

Machine Learning Engineer - I

Manthhan Software Pvt. Ltd.

Jul 2026 – Present · Jaipur

Fine-tuned SLMs and BERT-family models using LoRA, QLoRA, and PEFT; developed a RAG-based automated compliance checking system achieving 94% compliance detection accuracy across vendor bid documents; architected document workflows with hash-based deduplication, Prefect, MongoDB, AWS S3/MinIO; managed AWS SageMaker, MLflow, Git, Docker, quantisation, and model compression for production MLOps.

Data Science Intern

Manthhan Software Pvt. Ltd.

Jan 2026 – Jun 2026 · Jaipur

Designed and benchmarked document extraction pipelines across ML, NLP, and LLM approaches, improving average accuracy from 73% to 94%; built backend services with Flask, FastAPI, REST APIs, Docker, Prefect, MongoDB, MySQL, AWS S3/Bedrock, and pytest coverage.

AI Intern

Kenesis Labs Pvt. Ltd.

Jun 2025 – Nov 2025 · Chennai

Built multimodal pipelines using YOLO, CLIP, BLIP, and OpenCV for classification and duplicate detection; optimised PyTorch inference and reduced manual review workload by 40%.

Student Ambassador

Adobe

Events Organizer

IEEE

Student Partner

Internshala

1 yr 1 mo · Chennai, Tamil Nadu, India

Conducted outreach to 500+ people.

Lead

Coding Ninjas

1 yr 6 mos · India

Led student coding initiatives, organized learning sessions, and helped peers understand core programming and ML concepts. Guided beginners through doubts, shared resources, and built an active coding community on campus. Collaborated with the Coding Ninjas team to drive engagement and promote hands-on learning.

Student Placement Coordinator

Machine Learning Engineer

Manthhan

2 mos · Jaipur, Rajasthan, India

Data Science Intern

Manthhan

6 mos · Jaipur, Rajasthan, India

At Manthhan Software, I work on developing and fine tuning document extraction and classification systems for supply chain workflows using machine learning, deep learning, NLP, and LLM based approaches. A big part of my role involves building RAG pipelines, experimenting with LLM fine tuning, and training models that can handle real world document variability. I work closely with AWS S3, Bedrock, and MinIO for storage and model integration, and help connect ML systems to backend services through REST APIs. I also contribute to backend codebases, use Docker for deployment, and orchestrate workflows using Prefect. Beyond modeling, I focus on making systems production ready by writing unit tests and integration tests, debugging edge cases, and ensuring the pipelines remain stable and reliable in live environments.

Campus Ambassador

Adobe

9 mos · India

As an Adobe Student Ambassador, I represent Adobe on campus by encouraging students to explore creativity and digital storytelling through Adobe tools. This role gives me the opportunity to learn directly from Adobe experts, organize activities for my peers, and share ideas that connect technology with creativity. It is also helping me grow as a communicator and leader while being part of a global community of creative students.

AI Intern

Kenesis

6 mos · Chennai, Tamil Nadu, India

Worked on multimodal AI systems using YOLO, CLIP, BLIP, and PyTorch to improve image and video verification workflows. Built and optimized pipelines for audio, video, and vision data that reduced manual review effort by about forty percent. Also helped debug models, create evaluation setups, and guide two junior interns through training and deployment tasks.

Campus Ambassador

MyGov India

1 yr 7 mos · Āvadi, Tamil Nadu, India

Promoted national digital initiatives on campus and encouraged students to participate in government-led programs, competitions, and awareness drives. Helped with outreach, content sharing, and community engagement to increase student involvement in MyGov activities.