Cyprus Remote (Global) Full-time

Social Discovery Group is hiring a Senior NLP / LLM Engineer

About the Role

As a Senior NLP / LLM Engineer, you will drive the design and refinement of large language models to enhance the quality of conversational AI experiences. Your work will directly shape how users interact with social discovery products, ensuring responses are accurate, context-aware, and engaging.

What You'll Do

  • Experiment with LLM architectures and training methods—including supervised fine-tuning, RLHF, and adapter-based techniques—to improve model behavior and output quality
  • Build and maintain evaluation systems that measure performance using both offline benchmarks and live user feedback
  • Optimize models for fast, efficient inference in production, focusing on latency, throughput, and scalability
  • Adapt pre-trained models using LoRA, QLoRA, prompt tuning, and instruction tuning to meet specific functional needs
  • Work alongside data scientists, engineers, and product teams to integrate AI capabilities into product workflows
  • Improve the deployment stack by evaluating inference servers like vLLM and SGLang for better performance and reliability

Requirements

  • Proven experience fine-tuning LLMs using techniques such as SFT, LoRA, QLoRA, and prompt engineering
  • Hands-on background in training and modifying NLP models—not just API integration or RAG pipelines
  • Firm grasp of deep learning, transformer architectures, and standard NLP workflows
  • Familiarity with model evaluation, A/B testing, and metric-driven experimentation
  • Experience optimizing models for inference, including quantization and serving frameworks
  • Strong Python skills and proficiency with PyTorch, Hugging Face Transformers, and related ML libraries
  • Understanding of software engineering principles and production-grade development practices
  • Functional English communication skills (B2 or higher)

Benefits

  • Full-time remote work with the freedom to work from any approved location worldwide
  • 28 days of vacation and 7 wellness days annually for personal recovery or errands
  • Referral bonuses up to $5,000 for successful hires
  • 50% reimbursement for professional courses, certifications, and international conferences
  • Corporate discounts on English language training
  • Health benefits, including up to $1,000 per year in compensation for medical expenses for employees and family
  • Support for remote work setup: one-time reimbursement of up to $1,000 gross every three years for home office or co-working costs
  • Company-provided workspace in offices or co-working spaces, including desk, chair, and internet
  • Internal recognition program where peers award points redeemable for experiences, merchandise, and wellness perks
Required Skills
LLMsSFTRLHFLoRAQLoRAprompt tuninginstruction tuningtransformer architecturesNLP model developmentmodel evaluationfine-tuningvLLMDeep LearningMachine Learningclassical NLP LLMsSFTRLHFLoRAQLoRAprompt tuninginstruction tuningtransformer architecturesNLP model developmentmodel evaluationfine-tuningvLLMDeep LearningMachine Learningclassical NLP
Earn more as a remote developer

Performance pay that rewards your skills

Iglu's revenue-sharing model means top performers earn significantly more than traditional salaries. Choose your projects, deliver great work, and see it reflected in your pay.

Revenue-sharing compensation
Project choice & autonomy
International client base
Career growth support
Check compensation
Top earners exceed market rate
About company
Social Discovery Group

Social Discovery Group is the world's largest group of social discovery companies. It unites more than 40 brands, an investment fund, and a venture studio.

SDG solves the problem of loneliness, isolation, and disconnection, transforming virtual intimacy into the new normal.

The company serves 500 million users across 150 countries and connects people globally through social entertainment platforms.

All jobs at Social Discovery Group Visit website
Job Details
Department Machine Learning
Category data
Posted 10 days ago