Middle Prompt Engineer

Middle Prompt Engineer at Social Discovery Group

Social Discovery Group (SDG) is the 3rd largest social discovery company globally, uniting over 60 brands and serving 500 million users. Our mission is to combat loneliness, isolation, and disconnection by integrating virtual intimacy into daily life. Our diverse portfolio includes online communication platforms leveraging AI, game mechanics, and video streaming, such as Dating.com, DateMyAge, Cupid Media, Dil Mil, and Kiseki.

SDG actively invests in IT startups worldwide, with notable investments in Open AI, Patreon, Flo, Clubhouse, Woebot, Flure, Astry, Coursera, and Academia.edu.

We foster a collaborative environment of like-minded IT professionals dedicated to creating and enhancing globally impactful social discovery products. Our international team of 1200 professionals and digital nomads operates across the globe.

Our digital nomad teams work remotely from various locations including Cyprus, Malta, the USA, Armenia, Georgia, Kazakhstan, Montenegro, Poland, Latvia, Serbia, Spain, Portugal, UAE, Israel, Turkey, Thailand, Indonesia, Japan, Hong Kong, Australia, and many others.

Achieving Great Place to Work US Certification™ in August 2024 underscores our commitment to a workplace built on trust, pride, and camaraderie, valuing these above just perks.

We are seeking a Middle Prompt Engineer to join Cupid Media and significantly enhance the quality of our AI moderation systems.

In this role, you will manage high-volume user-generated content (over 100,000 photos and millions of messages daily) and continuously refine LLM prompts to automate the moderation of photos, profiles, and conversations.

Your Main Tasks Will Be

• Design and optimize prompts for LLMs used in moderation of user-generated content (UGC)
• Test and evaluate prompts across diverse scenarios to improve accuracy and reliability
• Analyze model outputs to identify hallucinations, errors, and moderation gaps
• Develop evaluation metrics and benchmarking pipelines (precision, false positives, false negatives)
• Create and maintain labeled datasets for evaluating moderation quality
• Run experiments with different LLMs to improve moderation performance
• Collaborate with engineering and moderation teams to improve AI moderation systems

We Expect From You

• 2+ years of experience working with LLM systems or applied AI
• Experience with prompt engineering in production environments
• Strong understanding of prompting techniques (zero-shot, few-shot, chain-of-thought, role prompting)
• Experience evaluating LLM outputs and analyzing errors (false positives / false negatives)
• Experience working with labeled datasets and evaluation workflows
• Basic programming skills (Python, JavaScript, or TypeScript)
• Familiarity with Git
• Analytical mindset and strong problem-solving skills
• Ability to work independently and take ownership of results

What Do We Offer

• REMOTE OPPORTUNITY to work full time
• 28 calendar days of vacation per year
• 7 wellness days per year (time off) for personal or household matters, without using sick leave
• Bonuses up to $5000 for successful applicant referrals
• 50% payment for professional training, international conferences, and meetings
• Corporate discount for English lessons
• Health benefits: Up to $1000 gross per year compensation for self-purchased health insurance or doctor’s fees for employees and close relatives if not eligible for corporate medical insurance
• Workplace organization: Equipped office/co-working space or reimbursement up to $1000 gross once every 3 years for home office setup/co-working rent
• Internal gamified gratitude system: Exchange colleague bonuses for merchandise, team building, massage certificates, etc.

Key skills/competency

• Prompt Engineering
• LLM Optimization
• AI Moderation
• User-Generated Content (UGC)
• Python
• Data Analysis
• Evaluation Metrics
• Problem-Solving
• Applied AI
• Git