Overview

Annapurna Labs, designs custom silicon powering AWS's cloud infrastructure. Custom SoCs live at the heart of Amazon ML servers - including Inferentia and Trainium Systems - delivering high-performance ML inference and training at cloud scale. We're looking for a Senior Manager, Physical Design Engineering to build and lead a world-class Physical Design team delivering multi-billion-transistor ML accelerator SoCs on latest tech nodes. This is a player-coach role: you will personally engage in the most critical technical decisions; chip-level floorplan architecture, multi-GHz timing convergence, high-speed I/O physical design, power delivery for KW class systems, and 2.5D / 3D integrations; while building a high-performing team and driving rigorous PPA targets, first-pass silicon success, and relentless continuous improvement. Key job responsibilities- Build, hire, and develop a physical design team of 10-15 engineers. Own headcount planning, performance management, career growth, and organizational scaling to support multiple concurrent tapeouts. - Own the physical design strategy and execution roadmap across Inferentia, Trainium, and future products: implementation architecture, hierarchical methodology, PPA target-setting, and technology node adoption. - Institutionalize continuous PPA improvement: benchmarking frameworks, regression tracking, design-quality dashboards, post-tapeout retrospectives, and structured improvement programs that compound across generations. - Provide hands-on technical leadership on the hardest problems: multi-GHz timing closure, PDN architecture, high-speed I/O physical design, and 2.5D / 3D cross-die integration. - Lead technology node enablement for 2nm and beyond: PDK evaluation, reference flow development, PPA pathfinding, and adoption recommendations. - Define and deploy innovative RTL2GDS methodologies and CAD flows. Champion AI/ML-augmented design automation (RL-placement, ML-guided ECO, predictive analytics). Optimize cloud infrastructure for PD compute. - Serve as part of senior technical interface with foundry partners and EDA vendors on tool roadmap, co-development, and technology co-optimization. - Own tapeout execution: readiness reviews, foundry submittal, mask data coordination, and post-silicon yield learning. Drive cross-functional alignment with RTL, DFT, STA, Package, and Validation teams. BASIC QUALIFICATIONS - 5+ years people management leading PD teams of 8+ engineers, including hiring, performance management, and organizational growth.- 10+ years hands-on ASIC/SoC physical design across multiple advanced nodes (7nm, 5nm, 3nm) with multiple successful tapeouts.- Technical depth across full RTL-to-GDSII: floorplanning, CTS, routing, timing closure, PDN design, IR-drop/EM signoff, physical verification, and tapeout.- Mastery of both EDA ecosystems: Synopsys (Fusion Compiler/ICC2, PrimeTime, StarRC, RedHawk-SC) AND Cadence (Innovus, Tempus, Quantus, Voltus). Physical verification with Siemens Calibre (DRC, LVS, PERC).- Deep expertise in advanced STA (MMMC, POCV/AOCV, SSTA, PBA, IR-aware timing) and power integrity (static/dynamic IR-drop, EM, rush current, Ldi/dt, package-aware analysis).- Proven expertise in high-speed PD (multi-GHz datapaths, SerDes, DDR5, HBM, PCIe Gen5+, UCIe) and low-power design (UPF, power gating, DVFS, retention).- Experience with hierarchical methodologies for very large SoCs (>1B transistors). Expert scripting in TCL, Python, Perl. PREFERRED QUALIFICATIONS - Experience leading PD for AI/ML accelerators, GPUs, or high-performance data-center SoCs at leading semiconductor companies or hyperscalers.- Direct 3nm GAA/Nanosheet experience; 2nm or sub-2nm engagement. Familiarity with CFET and backside power delivery (BSPDN).- Quantifiable track record of continuous PPA improvement and metric-driven flow development across multiple projects and nodes.- Strong program execution: managing concurrent tapeouts, milestone tracking, risk identification, and executive