🚀 Senior Machine Learning Engineer

Senior Machine Learning Engineer (Document Intelligence, Agentic AI & Operational Finance)

Location:

Italy (hybrid) ·

Employment:

Full-time ·

Seniority:

Mid/Senior

About the client (confidential)

Our client is a fast-growing enterprise SaaS company serving Italy’s

field-services

ecosystem with an operational platform spanning sites, finance, and administration, powered by

applied AI

(document & image understanding, accounting automations, chat-based assistant, cash-flow and margin forecasting). Full company details will be shared during the first interview.

Role mission (descriptive)

As an ML Engineer, you will lead the design, training, and productionization of

Document & Data Intelligence

solutions and enable

agentic AI

capabilities through an internal

AI SDK

that unifies model access, tool-calling, evaluation, and telemetry. You’ll work on high-impact use cases, structured extraction from

invoices/DDT/SAL

and field images,

bank↔document reconciliation ,

anomaly detection

on costs/margins, and

cash-flow/cost forecasting,

embedding models and agents into

Accounting, Finance, and Field Ops

workflows with strong security, GDPR-compliant data handling, and

MLOps/LLMOps

best practices (versioning, observability, feedback loops).

What you’ll do — example activities (non-exhaustive)

Design and maintain the

AI SDK

(Python): unified APIs, provider adapters, tool registry, JSON-schema/Pydantic validators, prompt/model versioning, evaluation harnesses, tracing/metrics, retry/backoff, and latency/cost budgets. Build end-to-end pipelines across

OCR/NLP/LLM , tabular ML, and

time series , with clear metrics and production guardrails. Develop

stateful agents

(planning, tool-use, memory, reflection) with guardrails and HITL; automate report/timesheet updates. Partner with Product/Backend/Frontend to embed models/agents into workflows (incl. banking integrations and Italian fiscal systems such as

SDI ). Run

AgentOps/MLOps : CI/CD for models & prompts, dataset/versioning, observability (traces/logs/tokens), review queues. Uphold

privacy, security & compliance : PII handling, RBAC/RLS, audit logs, data retention.

Minimum qualifications

3+ years (5+ for Senior) in applied ML with

production

delivery of OCR/NLP or tabular models. Strong data engineering (feature engineering, dataset curation, evaluation). Familiarity with

time series

and

anomaly detection . Proven experience building

developer SDKs/libraries

(API design, packaging, versioning, docs, testing, CI). Preferred experience with

TypeScript

(for SDK/frontend) Professional Italian and English. No visa sponsorship:

EU work authorization required and must already be living in the Milan area.

Nice to have

Agentic AI

(LangGraph/LangChain or equivalents), tool abstraction, memory stores (vector DB/Redis), structured outputs. Hugging Face

ecosystem (Transformers/Datasets/Evaluate, Hub/model cards, LoRA/PEFT, TRL); optional self-hosting (vLLM/Triton/TensorRT-LLM). RAG

over fiscal/ERP documents; hybrid search and metadata filters. Domain knowledge:

Italian invoices, DDT, SAL, bills of quantities , banking/fiscal flows (SDI).

Tech stack (indicative) Python · PyTorch/TensorFlow · scikit-learn/XGBoost · Transformers/spaCy · OCR/Document-AI libs · FastAPI · MLflow/W&B · vector/relational DB · vLLM/Triton · containers/K8s · OpenTelemetry.