Johan Ehrenfors

One of two co-founders of a B2B SaaS startup that came out of the Antler NOR7 residency. Together we have created a customer feedback analysis tool for product organizations to better understand what users are telling them at scale, based on unstructured data like reviews, NPS surveys, and support tickets. When paired with a contextual understanding of the business and the organization, insights are automatically routed to the right internal stakeholders, to ensure relevancy. Achievements include: - Securing an initial investment of 150k€ and additional grants - Kicking off 3/4 pilot projects with household-name Swedish scale-ups 📦👕🎛️ (+🛴) - Conducting discovery calls and user interviews, in order to build up and prioritize a healthy backlog - Building a B2B SaaS from scratch, including technologies such as Gemini LLMs, Google Cloud Platform, Qdrant vector DB, PostgreSQL relational DB, FastAPI-based REST API, NextJS web-app, SendGrid emails, etc. - Building LLM-heavy workflows to automatically generate and evaluate insights, based on ~100k new feedback comments every month - Building a tool-equipped agentic AI chatbot to naturally interact with the feedback data

Joined Antler's 3-month program for aspiring entrepreneurs looking to find co-founders and to validate startup ideas with expert guidance. I have since continued to the second phase, after securing an initial investment. Stay tuned!

Software engineer in the Merchant Experience team, working mainly with TypeScript and NodeJS to build the self-service Partner Portal, where merchants can administrate their orders being delivered through Instabee. Key contributions include: - Designing and building new solutions tailored for smaller merchants, rather than big enterprise ones - Moving to a new authentication and role-base authorization system, including a self-service user management system - Moving Porterbuddy (another Instabee brand active in Norway) merchants to the multi-brand Partner Portal, including support for new functionality that can benefit all brands

As part of the Platform team (officially "Platform Engineer"), I work on any software that enables us to create the Furhat robot, SDK, and other related solutions, as well as documentation and providing technical support to customers. Due to the small team size working on creating an entire physical product, my work spans across multiple tech stacks, including Kotlin/Java for the core Furhat SDK and robot logic, C++ (computer vision, servo control, firmware, etc), Javascript/TypeScript (React-based user interface, NodeJS+Express robot and cloud-based APIs), as well as utilizing AWS infrastructure, Gitlab CI, Docker, etc. Both the product and development environment is Linux-based. Close collaboration with colleagues across the company is necessary to meet business needs. Some of the more notable contributions have been work on integrating the new FaceCore facial animation system, Skill Library, improving the computer vision and neck servo-handling, and developing multi-language, multi-modal robot interactions for a large external project.

Organized material for and represented the company at student events, including presenting at a webinar for 500 students. I have also interviewed and selected thesis candidates, and was the industry supervisor for two thesis students.

As a junior consultant at Cybercom, I am part of the internal XFT Swift Radio Team working for Ericsson, where I am focusing on embedded systems development in C++ and tests in Java. I was the main driver for a service to detect interaction sequences that may degrade the hardware, which was presented to 200 employees. Additionally, I have also worked with keeping the CI/CD track open, power saving features, and refactoring tasks.

At Cybercom Kista, I am part of a two-person team with the end goal of getting a quadcopter to autonomously explore a limited area, in order to generate a map of the surroundings. Our academic contributions consist of improving an existing open source exploration method (NBV Planner) and adding adaptive velocity control to maximize the area that can be explored on a single battery charge. For this project, we are using technologies, software, and methods such as: - ROS middleware, with C++ modular software nodes - Intel RealSense RGB-D camera - SLAM (RTAB-MAP) - RRT path planning - Software-in-the-loop (SIL) simulations in Gazebo - Requirements based engineering - Modular software architecture

As a student consultant, my skills have been sold for early-stage prototype development purposes. This has included aspects of electronics, basic simulations, programming, and mechanical prototyping - including lots of independent solution finding.

I represent that Mechatronics Master's Program at KTH for ICES. My role involves voicing my opinion and raising issues at quarterly meetings, and tackling tasks with the goal of increasing collaboration between students, research, and the industry. One example of my contributions was the introduction of "Pizza with an Engineer" evenings where students get to meet engineers who work at ICES member companies on a face-to-face basis.

- Robust Mechatronics (MF2043): Preparing materials, guiding and examining students in labs with a focus on designing robust electronic PCBs with prototyping machines. - Bachelor Thesis in Mechatronics (MF133X): Providing assistance for a multitude of different student projects. - Basic Electrical Engineering (MF1016): Lab assistant and exam grader on a wide variety of topics within electrical engineering for Bachelor students.

Link to the official program description: http://goo.gl/uVfSw1