Gunjan Paul

SWE AI Engineer (Contract) at ValleyDAO

Built multi-agent knowledge graph and RAG systems for continuous change detection, fact validation, and contradiction detection across technology and business workflows. Architected and deployed fine-tuned Llama-based models for climate and synthetic biotechnology research, improving analysis quality. Implemented vLLM optimization reducing inference latency and increasing throughput. Designed scalable multi-agent LLM infrastructure with load balancing and dynamic resource allocation handling 20K+ daily requests. Created automated graph generation for production flow visualization and built an LLM-driven grant recommendation system to match projects with opportunities.

Student Researcher at Computer Vision Center

Working on multi-modal 2D to 3D Gaussian Splatting with prompting and hierarchical segmentation in collaboration with CVLab EPFL. Implemented graph-based dense and overlapped object counting using 3D Gaussian Splat. Generated synthetic 3D semantic/instance segmentation data for plant-related CV research using Blender. Developed robust registration for low-overlap point clouds.

Backend and MLOps Engineer at Alectio Inc.

Developed active learning algorithms to improve data efficiency in large-scale ML projects. Built secure APIs and CI/CD pipelines for ML model deployment and monitoring using Kubeflow. Designed data pipelines with Apache Airflow, Kafka, PySpark, and Flyte. Engineered architectures for a data-labeling marketplace and a custom Data Lake, reducing annotation storage costs. Created a cloud-agnostic microservice architecture and led a team of five across DevOps, backend, and ML engineers.

Research Intern at Indian Institute Of Technology, Mandi

Developed a Unity3D open-world simulation to collect human data for training agents in complex search-and-retrieval tasks. Trained and benchmarked RL and cognitive models (Actor-Critic) to mimic human behavior.

Mitac Global Research Intern at University of Alberta

Developed Learning Analytics tools to visualize student activities on online forums. Built an NLP model using BERT for cognitive analysis of student discussions.

Computer Vision Engineer Intern at HAMMOQ

Developed deep learning models and data pipelines to automate resale processes. Deployed models on GCP & AWS, integrated with iOS/Android apps. Built web scrapers for eBay, Amazon, and Facebook Marketplace.

Student Researcher at Computer Vision Center

Developing multi-modal 2D-to-3D Gaussian Splatting with language prompting and hierarchical segmentation, in collaboration with CVLab EPFL. Graph-based dense and overlapped object counting using 3D Gaussian Splat. Created synthetic 3D semantic and instance segmentation data using Blender for plant-related computer vision research. Implemented robust registration for low-overlap point clouds.

Backend and MLOps Engineer at Alectio Inc.

Developed Active Learning algorithms to improve data efficiency in large-scale ML projects. Built secure APIs and CI/CD pipelines for ML model deployment and monitoring using Kubeflow. Designed data pipelines using Apache Airflow, Kafka, PySpark, and Flyte. Engineered architectures for a data-labeling marketplace and a custom Data Lake, achieving substantial cost savings in annotation storage from $5.5K to $250. Delivered a cloud-agnostic microservice architecture for scalable deployment and led a team of five including DevOps, backend, and ML engineers.

Research Intern at Indian Institute Of Technology, Mandi

Developed Unity3D open-world simulation to collect human data for training agents in complex search-and-retrieval tasks. Trained and benchmarked RL and cognitive models (e.g., Actor-Critic) to mimic human behavior.