Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m an AI engineer with a background in AEC and proptech, most recently at Autility, where I developed and deployed deep learning and LLM-driven solutions that streamlined building workflows and reduced project timelines from weeks to hours. Passionate about applying AI to the built environment, I focus on turning complex technical ideas into practical tools that create measurable impact in construction, energy, and property technology.…I’m an AI engineer with a background in AEC and proptech, most recently at Autility, where I developed and deployed deep learning and LLM-driven solutions that streamlined building workflows and reduced project timelines from weeks to hours. Passionate about applying AI to the built environment, I focus on turning complex technical ideas into practical tools that create measurable impact in construction, energy, and property technology.

Sounok Sarkar

Developer, Full Stack Developer, Data Scientist, +3





I’m an AI engineer with a background in AEC and proptech, most recently at Autility, where I developed and deployed deep learning and LLM-driven solutions that streamlined building workflows and reduced project timelines from weeks to hours. Passionate about applying AI to the built environment, I focus on turning complex technical ideas into practical tools that create measurable impact in construction, energy, and property technology.…I’m an AI engineer with a background in AEC and proptech, most recently at Autility, where I developed and deployed deep learning and LLM-driven solutions that streamlined building workflows and reduced project timelines from weeks to hours. Passionate about applying AI to the built environment, I focus on turning complex technical ideas into practical tools that create measurable impact in construction, energy, and property technology.

Available to hire

I’m an AI engineer with a background in AEC and proptech, most recently at Autility, where I developed and deployed deep learning and LLM-driven solutions that streamlined building workflows and reduced project timelines from weeks to hours. Passionate about applying AI to the built environment, I focus on turning complex technical ideas into practical tools that create measurable impact in construction, energy, and property technology.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Work Experience

Software Engineer at Voyansi

April 1, 2023 - September 25, 2025

Developed a full-stack cross-platform web application acting as an App Store for BIM, enabling enterprises and collaborators to create organizations, perform custom user management, and do version control for Autodesk plugins. Improved client UX by reducing time spent managing plugins. Built a 3D model viewer (Three.js/WebGL), client dashboards, and interoperability plugins for BIM software applications.

Software Engineer (Machine Learning) at Autility

April 1, 2023 - Present

Implemented and evaluated deep learning models (Mask R-CNN and YOLO-vX) for custom object segmentation from client images. Built an end-to-end deployment pipeline using FastAPI and Google Cloud Platform, plus a dynamic frontend for interactive result modification. Reduced IFC building creation time from about a week to a few hours. Orchestrated RAG-based LLM workflows (LangChain, Vespa DB, Redis) within a Django platform, supporting 100+ users and 10k+ ingested documents; improved retrieval with semantic reranking and ensured reliability through Kubernetes monitoring. Contributed features include conversation summarizer, image descriptor, and guided incident detection from building management docs using the GPT-4 API; the incident detector reduced task time from over an hour to minutes and helped win pilot customers.

Software Engineer at Voyansi

April 1, 2023 - September 5, 2025

Developed a platform serving as an Appstore for BIM enabling enterprise user management and version control for Autodesk plugins. Delivered full-stack cross-platform web applications including a 3D model viewer with Three.js/WebGL and BIM interoperability plugins, enhancing project lifecycle tracking and client management capabilities.

Software Engineer (Machine Learning) at Autility

April 1, 2023 - Present

Implemented and evaluated deep learning models such as Mask RCNN and Yolo-vX for custom object segmentation, establishing end-to-end deployment pipelines with FAST API and Google Cloud Platform. Orchestrated RAG-based LLM workflows utilizing LangChain and Vespa DB for optimized document retrieval, including semantic reranking and observability enhancements. Contributed intelligent features like conversation summarizers, image descriptors, and guided incident detectors using GPT-4 API, significantly reducing user task times.

Technology Specialist/Developer at HOK

September 30, 2020 - September 5, 2025

Reduced drafting time by developing custom tooling for international projects, created Sketch-Gan app generating images from building sketches using Pix2Pix GAN architecture, leveraging web data scraping and image processing to improve outputs. Customized YOLO v3 object detection models trained on scraped datasets to identify buildings and trees in images. Won best hack award and supported firm-wide tooling adoption.

Technology Specialist/ Developer at HOK

September 1, 2020 - September 25, 2025

Reduced drafting time for an international project from 4 weeks to 8 minutes by developing custom tooling. Contributed to Sketch-Gan: an app that generates images from conceptual building sketches using Pix2Pix; built a pipeline pulling Flickr image data (SQL and Beautiful Soup) into a repository, then applied OpenCV-based image modification and tuned Pix2Pix hyperparameters. ~20 weekly active users and the Best Hack Award at a major AEC hackathon. Also customized a YOLO v3 object detector trained on labeled datasets (Google v5) to detect buildings and trees using Keras/TensorFlow.

Research Technician 1 at Digital Building Lab, Georgia Institute of Technology

December 1, 2018 - September 25, 2025

Developed a Python-based digital framework on software APIs for procedural generation of planar assembly building systems; implemented multiple building typologies based on user input using NumPy and SciPy, contributing to funding for subsequent lab work.

Research Technician 1 at Digital Building Lab, Georgia Institute of Technology

December 1, 2018 - September 5, 2025

Developed a digital framework for procedural generation of planar assembly building systems using Python and statistical methods. Implemented various building typologies that facilitated further research funding, supporting architectural and digital building research projects.