BConfident - Automated Video Interviewing with Emotion Detection
An automated video interviewing platform for iOS that analyzed candidates' emotional states in real-time during recorded responses. Using an emotion detection SDK from an Egyptian MIT scholar, the app could detect confidence, potential deception, excitement, and confusion while candidates answered interview questions. The platform aimed to provide deeper insights than written assessments by capturing authenticity, body language, and emotional state. A year-long fundraising journey culminated in a EUR 200K offer from a German research center, which was ultimately declined.
Problem Solved
Written assessments allowed candidates to prepare rehearsed answers that hid their authentic reactions, while traditional video interviews lacked objective emotional analysis
- •Built iOS application with video recording and playback capabilities
- •Integrated emotion detection SDK with calibration for different demographics
- •Designed company and candidate user flows
- •Implemented Firebase backend for interview management
- •Led fundraising efforts across Egypt, Germany, and Latvia
- ✓Built automated video interviewing platform with real-time emotion detection
- ✓Integrated ML SDK capable of detecting confidence, deception indicators, excitement, and confusion
- ✓Partnered with Egyptian MIT scholar who provided emotion detection SDK (free until commercial)
- ✓Designed dual user flows for companies (create questions, review responses) and candidates (record video answers)
- ✓Implemented data collection mechanism for training emotion models on different demographics
- ✓Added multilingual transcription support for global market
- ✓Conducted year-long fundraising journey across multiple countries
- ✓Received EUR 200K investment offer from German research center with VC connections
Technology Stack
Challenge
Emotion detection ML needed calibration for different demographics and lighting conditions to provide reliable analysis across diverse candidates
Solution
Built data collection mechanism within the app to gather training data and worked with SDK creator to improve model accuracy across different demographics
Impact
Improved emotion detection reliability for varied user populations
Challenge
Egyptian angel investors and VCs offered amounts too small to cover operational salaries, making local fundraising unviable
Solution
Expanded fundraising search internationally, ultimately connecting with German research center that had VC partnerships and offered significant investment
Impact
Received viable EUR 200K offer after exploring multiple funding sources
Challenge
Best investment offer required relocating to Germany, rebranding the product, and working as long-term interns - incompatible with established life in Latvia
Solution
Difficult decision to decline the offer rather than uproot established life and accept unfavorable terms
Impact
Project discontinued despite receiving real investor interest and funding offer
Situation
Written assessments in hiring allowed candidates to prepare polished, rehearsed answers that masked their authentic reactions. While video interviews captured more information, reviewers lacked objective tools to analyze candidate emotional states. Companies like HireVue and Unilever were exploring similar concepts, indicating market interest.
Task
Build an automated video interviewing platform that would use machine learning to analyze candidate emotions in real-time, providing recruiters with objective insights into confidence, authenticity, and engagement.
Action
Khaled built an iOS application with two distinct user flows: companies could register, create interview questions, and review video responses with emotion analysis overlays; candidates would enter interview codes and record video answers while the ML analyzed their emotional state. He partnered with an Egyptian MIT scholar who had developed an emotion detection SDK, integrating it with calibration for different demographics. The team spent a year fundraising, starting with Egyptian angels and VCs who offered insufficient amounts, then expanding internationally. A German research center with VC connections offered EUR 200K, but with conditions: relocating to Germany, rebranding under the research center, and working as long-term interns.
Result
The fundraising journey validated genuine investor interest in the concept, culminating in a significant EUR 200K offer. However, the conditions required sacrificing an established life in Latvia for unfavorable terms. Khaled declined the offer, acknowledging he wasn't brave enough or fully committed enough to make that sacrifice. The project was discontinued, leaving lessons about the importance of timing, full commitment, and the hard decisions entrepreneurs face between opportunity and life circumstances.
Technical
- • iOS video recording and playback implementation
- • ML SDK integration and calibration
- • Firebase backend for multimedia applications
- • Real-time video processing on mobile devices
Soft Skills
- • Startup fundraising across multiple countries
- • Investor pitch development and presentation
- • Difficult decisions about personal sacrifice vs business opportunity
- • Partnership building with research institutions
Key Insights
- 💡 Timing and full commitment matter more than the idea itself
- 💡 Investment offers often come with conditions that fundamentally change the venture
- 💡 Prototype stage isn't ready to sell to enterprise B2B customers
- 💡 Life circumstances affect ability to fully commit to entrepreneurial ventures