СербияSocial Discovery Group
Machine Learning Lead
1 нед. назад
СербияLeadRemote
pythontransformer architecturellmtrainingfine-tuningevaluationsftlora
Social Discovery Group is a global social discovery company developing dating and communication platforms. The role involves leading a team of ML engineers and owning the ML strategy for dialogue systems with remote work opportunities.
About the company
- Social Discovery Group (SDG) is a group of social discovery companies.
- SDG solves the problems of loneliness, isolation, and disconnection - transforming virtual intimacy into the new normal.
- SDG’s products redefine the way people interact and connect with one another.
- Our portfolio includes social entertainment platforms designed to connect people online across different cultures and regions of the world.
- We bring together a team of like-minded people and IT professionals who specialize in creating and developing globally impactful social discovery products.
- Our international team of digital nomads works remotely from all over the world.
- We’re proud to be a two-time “Great Place to Work” winner (USA & Japan, 2024–2025) and a Top-5 Company for Work-From-Anywhere Jobs (FlexJobs, 2025).
Responsibilities
- Own the ML strategy for dialogue systems: decide what to build and in what order, and tie those bets to business metrics (ARPU, retention, chat depth).
- Lead a team of 3 ML engineers — set the technical bar, distribute work, hire and let go, grow the people you keep.
- Drive LLM post-training and model adaptation end to end: SFT, LoRA/QLoRA, preference optimization (DPO / ORPO / SimPO / GRPO), dataset construction.
- Design agent harnesses and tool-using LLM systems: tool calling, structured outputs, routing, retries, memory and context, guardrails.
- Build the evaluation layer: offline and pairwise evals, LLM-as-a-judge, regression suites that actually correlate with A/B outcomes.
- Cut dialogue failure modes — loops, contradictions, persona drift, context loss, generic replies — and keep inference efficient on latency and cost per message.
- Stay hands-on where it matters (roughly 10–20% of your time): prototypes, debugging agent traces, reviewing your team's work.
Requirements
- Technical degree and a real ML engineering background — you have trained and shipped models yourself, not only managed people who do.
- Experience leading a team of 3–5 ML engineers, including hiring and performance decisions.
- Expert-level Python, solid understanding of transformer architecture and modern LLM behavior, hands-on with training, fine-tuning and evaluation.
- Production experience with LLM systems: APIs, streaming, batching, fallbacks, cost/latency trade-offs, observability over traces and transcripts.
- Ability to read fresh research and turn it into a prototype, an eval and a shipped change with measurable impact.
- Fluent Russian, ready to work in CET (±2) hours.
Nice to have
- Experience beyond text (computer vision, image generation, multimodal).
- Experience with long-running conversations and character consistency.
- Experience with vLLM / TGI / SGLang, DeepSpeed / FSDP / Accelerate.
- Experience with quantization, safety classifiers.
Conditions
- REMOTE OPPORTUNITY to work full-time.
- The initial pay level or pay range for this role will be shared with candidates during the recruitment process and before the commencement of employment.
- Vacation 28 calendar days per year.
- 7 wellness days per year (time off) that can be used to deal with household issues, to lie down and recover without taking sick leave.
- Bonuses up to $5000 for recommending successful applicants for positions in the company.
- 50% payment for professional training, international conferences, and meetings.
- Corporate discount for English lessons.
- Health benefits. Compensation up to $1,000 gross per year per employee for health insurance or doctor’s fees.
- Workplace organization: equipped workplace & equipment in offices or coworking, or reimbursement up to $1000 gross once every 3 years for workplace costs.
- Internal gamified gratitude system: receive bonuses from colleagues and exchange them for merchandise, team building activities, massage certificates, etc.