Quadric

Physical Design Flow & Methodology Engineer

Quadric  •  Pune, IN (Onsite)  •  6 hours ago
Apply
AI can make mistakes so check important info. Chat history is never stored.
66
AI Success™

Job Description

Quadric delivers its GPNPU as soft IP — RTL and implementation collateral — enabling customers to integrate our processor into their own SoCs across a range of process nodes and foundries. You will drive PPA optimization across IP configurations, build the scalable reference flows customers use to evaluate and integrate our IP, and provide hands-on implementation support to customers working toward their tapeouts.

Responsibilities

PPA Optimization & Analysis

  • Drive PPA analysis and optimization for Quadric GPNPU soft IP across process nodes and hardware configurations — timing, area, leakage, and dynamic power
  • Apply low-power techniques (clock gating, multi-Vt, operand isolation) and synthesis/P&R knobs to hit frequency and area targets
  • Characterize the IP design space across configurations and build PPA models that support customer evaluations and pre-sales engagements
  • Partner with RTL and architecture teams early to quantify tradeoffs and influence design decisions before they become costly to reverse

Reference Flow Development

  • Build and maintain a scalable RTL-to-GDS reference flow for Quadric soft IP that customers can use to evaluate, integrate, and close PPA in their own SoC environment
  • Ensure the flow is portable across supported process nodes with clear BKMs, SDC templates, floorplan scripts, and integration guidelines
  • Develop TCL and Python automation — and leverage AI coding tools such as Claude — to accelerate flow development, reduce manual effort, and improve repeatability
  • Qualify EDA tool updates and benchmark QoR impact before rolling into the reference flow

Customer Integration & Tapeout Support

  • Act as the primary PD contact for customers integrating Quadric soft IP, guiding them from evaluation through their SoC tapeout
  • Help customers adapt the reference flow to their process node, foundry PDK, and internal design environment
  • Triage and resolve customer-reported implementation issues — timing, congestion, power, or flow failures — working with internal teams to deliver fixes or updated collateral
  • Support FAE and business development with PPA feasibility studies for new customer engagements

Collaboration & Documentation

  • Work with architecture, RTL, and software teams to ensure IP deliverables meet customer-facing PPA targets
  • Document methodologies, BKMs, and optimization learnings; maintain process node bring-up guidelines to support IP portability

Requirements

Education & Experience

  • BS/MS in Electrical Engineering, Computer Engineering, or related field
  • 4+ years of ASIC or processor IP physical design experience focused on PPA optimization and flow development across advanced nodes

Technical Skills

  • Proficiency with industry-standard physical design tools from Synopsys or Cadence (synthesis, place-and-route, and timing analysis)
  • Experience with advanced FinFET process nodes (16nm and below); multi-node experience preferred
  • Strong TCL scripting and Python automation skills
  • Solid understanding of synthesis and P&R levers for PPA — timing paths, cell selection, congestion, and power intent
  • Hands-on experience with low-power design techniques and MCMM timing analysis
  • Comfort using AI tools (e.g., Claude, Copilot) to accelerate script development, automate repetitive EDA tasks, and improve workflow productivity
  • Understanding of DFT concepts (scan, ICG bypass) and their physical design implications

Nice to Have

  • Experience delivering soft IP to external customers or supporting SoC integrators through tapeout
  • Background in AI accelerator, NPU, or DSP processor IP implementation
  • Exposure to metrics-driven QoR tracking and large-scale synthesis run management
Quadric

About Quadric

Quadric licenses an AI processor architecture optimized for on-device inference. Only the Quadric Chimera GPNPU (general purpose neural processing unit) delivers high AI/ML inference performance while also running C++ code without forcing the developer to artificially partition application code between two or three different processors. Quadric's Chimera GPNPU processor IP core scales from 1 to 864 TOPS and seamlessly intermixes scalar, vector and matrix code.

Industry
Hardware & Semiconductors
Company Size
51-200 employees
Headquarters
Burlingame, CA
Year Founded
2017
Social Media