# Dhwani Trivedi > MTS @ Easybee AI | Building Agents | Prev @ MIT, ITU, ICTP | GenAI | Data Science | MLOPS | Python | Machine Learning | SQL Location: San Francisco Bay Area, United States Profile: https://flows.cv/dhwanitrivedi Feel free to reach out at dhwanitrivedi999@gmail.com A Machine Learning Engineer with a Master's in AI Machine Learning and Research (2023), I possess a blend of academic and practical experience, notably in scaling ML models for marine monitoring and advancing applied science for product development at MIT. My project achievements include end-to-end novel recipe recommendation engine that improved e-commerce precision by 40%, an AI self-diagnosis tool recognized by Harvard Medical School, and an AI-powered Android application for weight estimation showcased at UNESCO MGIEP. With expertise in enhancing performance, reducing latency, and deploying AI products, I am eager to transition into industry roles where I can drive innovation in machine learning and data science. I have also given talks at Google as WomenTechmakers Ambassador, UN, UNESCO, ITU, ICTP, Harvard Medical School, UNESCO MGIEP on AI for Social Good. I have skills in PyTorch, Tensorflow, Python, SQL, Tableau and AWS, Azure and GCP. Apart from this I am also a UN SDGs and Policymaking enthusiast ## Work Experience ### Forward Deployed Engineer @ EasyBee AI Jan 2025 – Present | Cambridge, Massachusetts, United States - Deployed core platform, a multi-tenant conversational and voice AI platform automating customer service for storage facility operators handling unit availability, reservations, and booking inquiries via chat and voice (LiveKit + Deepgram). - Engineered LLM-driven intent classification with RAG-based knowledge retrieval and real-time tool execution, enabling agents to resolve customer requests end-to-end without human intervention. - Built production observability across Langfuse, Maxim AI, and CloudWatch with automated KPI tracking cost, latency, task completion rate, and adversarial robustness ensuring SLA compliance at scale. - Architected on AWS with PostgreSQL checkpointing and S3 Vectors, supporting per-tenant isolation, configurable business profiles, and seamless multi-business deployment from a single codebase. ### Member of Technical Staff @ EasyBee AI Jan 2025 – Jan 2025 Scaling web, voice and deep research agents and building other cool stuff ### Member @ CIC (Cambridge Innovation Center) Jan 2025 – Present | Cambridge, Massachusetts, United States Meeting people ### Adjunct Faculty @ Northeastern University Jan 2024 – Jan 2025 | Boston, Massachusetts, United States Teaching EAI6020(Infra ML/ML-Ops), ALY6000(Intro to Analytics) ### Machine Learning Researcher @ Northeastern University Jan 2023 – Jan 2024 | Boston, MA • Spearheaded the development of a PySpark pipeline to enhance Fast-RCNN model performance for fish detection, resulting in a 40% increase in precision. • Transformed 1TB of data from 500 video files into a structured CSV format, optimizing data flow and boosting training efficiency by 50% for fish surveillance initiatives. ### Research Engineer @ Aware, Inc. Jan 2024 – Jan 2025 | Boston, Massachusetts, United States - Led the development of an internal tool using transformer-based architectures, automating 80% of manual tasks and optimizing processes, resulting in a 30% increase in operational efficiency. - Managed a scalable database pipeline for training large language models (LLM), handling over 5TB of data for preprocessing, validation, and training to ensure model accuracy and relevance. - Developed an evaluation framework by implementing LLM-as-a-judge to automate the classification and routing of customer support tickets. - Reduced response time by 40% and improved resolution accuracy by 25%, enabling the team to handle 1.5 million support tickets annually with enhanced efficiency. ### Applied Scientist @ Massachusetts Institute of Technology Jan 2022 – Jan 2022 | Cambridge, Massachusetts, United States Project was funded by one of the FAANG companies. - Implemented Two-Bounce Light principles from "Imaging Behind Occluders" paper to 3D reconstruct shadow frames, bridging theoretical concepts to a practical application for enhanced visual representation. - Managed a cross-functional team and stakeholder communication and over saw product development to inception to implementation. - Prototyped native Swift App for accessing depth maps and capturing point clouds using iPhone’s LiDAR sensor. - Generated a 3D reconstruction of occluded object using shadows and two bounce light using ARKiT, Cocoa Touch and CoreML packages ### Program Manager @ International Telecommunication Union Jan 2021 – Jan 2022 | Switzerland ### Data Scientist @ Bhaskaracharya Institute For Space Applications and Geo-Informatics Jan 2021 – Jan 2021 | India -Successfully developed a deep learning model with 87% accuracy to perform land cover classification for change detection. - Two satellite images of the specific time and coordinates are cropped from Sentinel Hub for data collection. - Pre-processed using the GDAL library for dataset preparation. - Created mask images for semantic segmentation using LabelMe annotation tool on the training data set and trained the mask images using U-Net Image Segmentation model. -Tested using the image of the different periods if any change was detected change maps were displayed. ### Undergraduate Research Fellow @ Abdus Salam International Centre for Theoretical Physics (ICTP) Jan 2019 – Jan 2019 | Trieste, Friuli-Venezia Giulia, Italy -Selected among 12 out of 100 participants around the world and received a full travel grant from ITU to study about LoRa Communication technology for Machine Learning and IoT applications such as Ocean Monitoring, Weather Monitoring in the Workshop on Rapid Prototyping of Internet of Things Solutions for Science | (smr 3268). -Pitched an idea on AI Enabled Traffic Monitoring System using LoRaWAN during case study presentation ## Education ### Phd Dropout in Data Analytics Engineering- Thesis Track Northeastern University ### Bachelor of Technology - BTech in Electronics and Communication Engineering Birla Vishvakarma Mahavidyalaya ### Fellowship in Internet of Things Abdus Salam International Centre for Theoretical Physics (ICTP) ## Contact & Social - LinkedIn: https://linkedin.com/in/dhwanitrivedi9 - Portfolio: https://dhwanit99.github.io/ --- Source: https://flows.cv/dhwanitrivedi JSON Resume: https://flows.cv/dhwanitrivedi/resume.json Last updated: 2026-04-11