Antonino Grasso

Board-sponsored AI initiative at a leading robotics manufacturer, built for its sales organization.

• Tech lead of a four-person team (PM, frontend, second backend engineer); took the project from my own pre-sales solution design to a working PoC in ~2 months, on schedule, and presented it to the client's board and executive management
• Designed and built an AI account-research pipeline in Kotlin (ontology-driven scoring, Anthropic API with web research, two-tier model routing at ~$1 per account) that cuts hours of manual research to a ranked assessment in ~7 minutes; architected for scale from the 78-account PoC toward ~15,000 accounts
• Built evaluation and observability into the pipeline: a sales-confirmed golden set with recorded LLM responses for zero-cost regression tests, and end-to-end tracing from raw account facts to the final assessment, so every score is explainable and auditable
• Now leading the move to production as tech lead: hardening the pipeline and scaling it from the PoC toward global availability across the client's sales organization

• Co-leading Device Insight's Competence Center AI, driving the AI Systems track — shaping how the organization approaches AI-system projects
• Built and own an agentic AI booking-check automation on Azure (Python) that flags mis-booked time in Harvest; in active use by PMs and productive with minimal maintenance, it has recovered ~€6k of revenue on ~€40 of run cost (~150x)

For a leading robotics manufacturer's CTO Office — intent-based robotics R&D, embedded AI, multi-agent orchestration.

• Lead R&D engineer for the agentic-orchestration layer of an intent-based robotics platform: scoped open CTO-Office briefs, built the prototypes, and drove the architecture proposals and platform-evolution decisions, coordinating with stakeholders across related focus topics
• Built a working multi-agent orchestration prototype in Python, running on NVIDIA Jetson — evaluated LangGraph and Microsoft Agent Framework hands-on, picked direction, prototyped on-device inference (vLLM, quantization, with measurable token-throughput gains) designed to run offline / fully on-robot, plus error-tolerant task execution
• Built ROS2-MCP interop between the robotics stack and the agentic orchestration layer; validated it end-to-end on real hardware by wiring a five-finger robotic hand's gesture actions through ROS2 up to MCP
• Authored an 80-page internal whitepaper that drove platform-evolution decisions — selected which components advance to next-stage research, which architectural directions the platform takes

Cloud-native SaaS — collaborative engineering platform on AWS, shipped to 10+ enterprise tenants.

• Designed and shipped the token-based licensing system end-to-end — product requirements, database schemas and Flyway migrations, observability instrumentation, and dev/stage/prod rollout. Live with 10+ enterprise tenants.
• Built the add-on management feature (S3 upload, association, lifecycle state changes, deletion) and integrated cloud-based software licensing
• Owned platform-wide observability — instrumented backend services with OpenTelemetry, set up tracing and metrics, built the local observability stack used across both dev teams
• Worked across backend, AWS infrastructure (AppStream, ECS), and DevOps — generalist scope on a 3–5 engineer team, one of two delivery teams in a ~20-person product org
• Collaborated closely with PMs, POs, and Go-to-Market on prioritization and product direction; represented the product at Automatica 2025 with live customer demos

Dual study programme — alternating university semesters and full-time engineering rotations across IoT projects.

• Designed an AWS Cognito-based auth/authz system for an IoT product line — shipped during the dual-study period and still running in production today
• Designed and set up the Domain-API end-to-end testing framework for a wind-energy client — defined the testing strategy spanning domain API down through simulated devices to the data source
• Rotated across HVAC, access control, and green-energy IoT projects — built early breadth across production distributed systems, end-to-end lifecycle exposure (architecture → engineering → operations), and fast onboarding into unfamiliar contexts