Senior AI Backend Engineer (Product Neuro Forge Team)

AppFollow is an App review management and ASO platform.
AppFollow is an app review management and ASO platform. Our main goal is to ease the everyday routines of app developers, product managers, marketing teams, customer support, etc. AppFollow helps you gather and manage your apps and games data, increase app average rating, improve app store rankings, and app user loyalty.
Ratings and reviews are our core data, and AI is how we turn them into value for our customers, helping them automate routine work with user feedback and save time: feedback categorization, review summarization, AI-generated replies, semantic search, anomaly detection, and conversational insights.
This fully remote role is for a Senior AI Backend Engineer who will drive these capabilities end-to-end, from prototyping to production.

You'll work with both commercial LLMs and open-source models, build evaluation pipelines, and ship ML-powered features used worldwide by app and game teams, as well as anyone working with digital user feedback. This is an applied engineering role at heart: your models won't stay in notebooks. Together with our backend team, you'll also develop the product functionality, services, and APIs around them. It's not a research or data science role: you won't train models from scratch, and fine-tuning experience is a plus, not a must.

About the Role
Own AI/ML features end-to-end: prototype, production, monitoring and iteration
Design and build LLM-powered features on top of reviews and ratings data: feedback categorization, summarization, reply generation, semantic search, anomaly detection, conversational and agentic scenarios
Develop and optimize AppFollow product functionality beyond ML: services and internal and public APIs to core data entities (apps, reviews, ratings)
Contribute with other developers to the overall system architecture; design, build, and document an efficient, testable, reliable, secure, and scalable codebase
Develop RAG and semantic search capabilities: embeddings, vector storage, retrieval quality
Optimize quality, latency, and cost of LLM inference in production
Build and maintain pipelines for quality evaluation: datasets, metrics, offline evals, LLM-as-a-judge, A/B tests; where it pays off, also for model training and fine-tuning
Work with commercial LLM APIs (OpenAI, Anthropic, Google) as well as open-source models (Llama, Mistral, Qwen, etc.): model selection, adaptation, fine-tuning, and deployment
Integrate new features with the existing system, including legacy services, and collaborate with the Platform team to deploy to various environments and ensure smooth operation
Track state-of-the-art in NLP/LLM, run experiments and POCs, and turn the promising ones into product features

About you
5+ years of software development experience; strong production Python (asyncio)
Proficiency in FastAPI for API development
Strong SQL skills (MySQL or PostgreSQL), experience with ORM frameworks (preferably SQLAlchemy)
Experience with unit testing (pytest)
3+ years of hands-on ML/NLP experience with models shipped to production; integrating, serving, and evaluating models counts, not only training them
Practical experience with LLMs: prompt engineering, RAG, working with both commercial APIs and self-hosted models
Experience building model quality evaluation processes: metrics, eval datasets and pipelines, A/B testing
Working knowledge of PyTorch and the Hugging Face ecosystem (transformers, datasets)
Upper-intermediate English or higher

It would be nice to have
Experience fine-tuning open-source models (LoRA/PEFT)
Experience serving open-source LLMs in production (vLLM, TGI, Triton) and working with GPU infrastructure
Experience with vector databases (e.g. pgvector)
Experience with agentic and orchestration frameworks (LangChain, LangGraph) and eval/observability tooling (MLflow, Langfuse)
Experience with data processing pipelines and automation (e.g. Airflow, Prefect)
Experience with cloud-based services (AWS), NoSQL databases (MongoDB), message brokers (RabbitMQ, Kafka)
Classical ML/NLP background (text classification, clustering, topic modeling)
Open-source contributions, publications, or pet ML projects you're proud of
Benefits we offer
Full-time remote job.

Though you're always welcome to spend time with us in monthly All hands in our hubs: Helsinki, Belgrade, Tbilisi, Batumi, Yerevan
Paid Vacation and Sick leaves. Take the time you need to stay motivated, charged, and balanced. By prior agreement, you can have days off for special occasions
Generous social benefits package including health insurance, equipment reimbursement, home office moderation bonus, and many more
Stock options bonus according to the employee stock ownership plan
You'll have executive-level visibility into how the company is run and performing. We are always ready to provide dedicated support and fast-track your onboarding, including giving you the tools you need to be successful.

The biggest benefit is our awesome AppFollow team. We're a team of open-minded and friendly high-skilled professionals that enjoy creating a great product, growing together, and supporting each other.
Jump on the board!

Hiring process
HR screening interview — 15 min
Backend Technical interview — 90 min
AI Engineering interview — 60 min
Culture fit interview — 60 min
Recommendations check
Expected timeline: 2–4 weeks from application to offer.

Hint
Want to increase your chances? Please ensure your LinkedIn profile is complete, up to date, and accurately reflects your experience before submitting your application.