Senior Applied ML Scientist – Search & Recommendation

Comparables.ai; Deutschland; Hybrid; EUR 90.000 - 150.000; Vollzeit

Zusammenfassung

Comparables.ai is seeking a Senior Applied ML Scientist to design, implement, and scale high-performance hybrid retrieval and ranking systems for financial discovery workflows. You will lead the development of hybrid, LLM-based retrieval pipelines and contribute to productionized ML solutions across large-scale datasets.

The role emphasizes applying transformer-based methods, domain adaptation, and efficient retrieval to deliver accurate company similarity and valuation benchmarking at global

Qualifikationen

• 5+ years of applied ML experience in production search or recommendation systems.
• Experience with hybrid retrieval architectures and both sparse and dense retrieval.
• Experience fine-tuning LLMs for retrieval and ranking tasks.
• Experience with domain-specific LLM adaptations.
• Experience with large-scale textual datasets.

Aufgaben

• Design and optimize hybrid search systems combining lexical, sparse, and dense retrieval.
• Build and deploy LLM-based retrieval and ranking models.
• Fine-tune transformer and LLM architectures for retrieval tasks.
• Develop novel LLM-based retrieval and ranking approaches beyond embeddings.
• Design recommendation strategies for company similarity, valuation comparables, and market workflows.
• Improve embedding representations for structured and unstructured data.
• Define and optimize relevance metrics tied to product outcomes.
• Run offline evaluations and online experiments.
• Balance quality, latency, and cost in production systems.
• Collaborate with backend and AI engineers to productionize models.

Kenntnisse

Hybrid retrieval

Sparse retrieval

Dense retrieval

Transformer models

LLM fine-tuning

Python

PyTorch

HuggingFace

Sentence-Transformers

PEFT LoRA

Vector databases

Tevatron

FlagEmbedding

Tools

Tevatron

FlagEmbedding

Jobbeschreibung

Senior Applied ML Scientist – Search & Recommendation

We are looking for a senior applied ML scientist to join our team.

Location: Helsinki Onsite (Preferable), Finland & EU Remote / Hybrid

The Opportunity

Comparables.ai builds the intelligence infrastructure for financial decision-making.

We operate a dataset covering 350M+ companies and power company discovery, peer identification, valuation benchmarking, and market research workflows. Search and recommendation quality are core to our product.

We’re hiring a Senior Applied ML Scientist to design, fine-tune, and scale high-performance hybrid retrieval and ranking systems.

What You’ll Work On

• Design and optimize hybrid search systems combining:
  • Lexical retrieval
  • Sparse representations
  • Dense embedding-based retrieval
• Build and deploy LLM-based retrieval and ranking models.
• Fine-tune transformer and LLM architectures for:
  • Retrieval & Ranking tasks
  • Financial domain adaptation
• Develop novel LLM-based retrieval & ranking approaches, beyond off-the-shelf embeddings.
• Design recommendation strategies for:
  • Company similarity
  • Valuation comparables
  • Industry benchmarking
  • Market intelligence workflows
• Improve embedding representations for structured and unstructured financial data.
• Define and optimize relevance metrics (Recall@K, Precision@K, NDCG) tied directly to product outcomes.
• Run structured offline evaluations and online experiments.
• Balance quality, latency, and cost in large-scale production systems.
• Collaborate closely with backend and AI engineers to productionize models reliably.

What We’re Looking For

• 5+ years of applied ML experience in production search, ranking, or recommendation systems.
• Strong expertise in:
  • Hybrid retrieval architectures
  • Sparse and dense retrieval systems
  • Transformer-based and LLM-based retrieval and ranking tasks
  • Fine-tuning LLMs for retrieval and ranking tasks
• Experience building or adapting LLMs for domain-specific applications.
• Experience working with large-scale textual datasets.
• Strong Python skills and hands-on experience with PyTorch, HuggingFace, Sentence-Transformers, PEFT (LoRA), Tevatron, FlagEmbedding
• Proven experience shipping ML systems that impact user-facing metrics.
• Solid understanding of system constraints:
  • Scalability
  • Indexing strategies
  • Inference latency
  • Cost optimization

Strong Plus

• Experience in financial or structured knowledge domains.
• Experience with Elasticsearch and vector databases (Weaviate, Milvus, Pinecone, Vespa).
• Experience building custom cross-encoders or re-ranking models.
• Experience running large-scale A/B experiments.

Why This Role Matters

Search and recommendation quality directly determine the value of our product. You will shape how financial professionals discover companies, benchmark valuations, and analyze markets — at global scale.

• Massive proprietary dataset (350M+ companies)
• High-impact ML problems in a real-world financial domain
• Strong technical ownership
• Strong growth opportunity in a rapidly scaling AI SaaS company.