Anuj Bhattacharya FPGA Engineer
Santa Clara, CA • fpgaeng@gmail.com • +1 408-555-0153
Profile Summary
- FPGA Engineer with 9 years of experience designing high-speed network switch ASICs and FPGAs across datacenter networking, high-speed I/O, and packet acceleration, specializing in high-speed RTL design, UVM verification, and timing closure on large FPGAs.
- Solid technical background across FPGA vendors (AMD Xilinx, Intel Altera), HDL languages (SystemVerilog, VHDL), verification on UVM, toolchain (Vivado, QuestaSim), high-speed buses (PCIe Gen4, 100GbE), and FPGA SoC integration on Versal Premium, with strong fundamentals in constraint-driven design, coverage-driven verification, and disciplined timing closure on every release.
- Deep expertise in high-speed RTL design and HDL development, FPGA microarchitecture and resource partitioning, coverage-driven verification and formal sign-off, and timing closure on multi-million-LUT designs, applying practices such as UVM and SystemVerilog assertions and CDC analysis and clock-aware design to deliver high-throughput, timing-clean, and verifiable FPGA designs.
- Engaged collaborator working cross-functionally with ASIC, Hardware, Embedded Software, and Verification leadership in tape-out-paced silicon programs, contributing to RTL + verification-plan reviews, timing-closure war rooms, and post-bring-up retrospectives with a pragmatic, ownership-first mindset.
- Senior engineer who shares technical excellence and fosters a culture of design-review rigor and written verification-plan discipline through code-review coaching and verification-plan mentorship, while running FPGA design-review board sessions and authoring widely adopted RTL-style and UVM-pattern handbooks.
Technical Skills
- HDL & RTL:
- SystemVerilog, VHDL, Verilog-2001, HLS (C / C++), SystemVerilog assertions (SVA), Chisel (intro)
- FPGA Platforms:
- AMD Xilinx (UltraScale+, Versal, Zynq), Intel Altera (Stratix, Agilex, Cyclone), Lattice, Microchip Microsemi
- Verification:
- UVM, cocotb, SystemVerilog assertions, OVM legacy, formal verification (JasperGold, OneSpin), constrained-random testbenches
- Synthesis & P&R:
- Vivado, Vitis, Quartus Prime, Libero SoC, XDC / SDC constraints, timing analysis, design partitioning, Design Compiler
- Clocking & CDC:
- MMCM / PLL, async FIFOs, gray-code, multi-clock design, Conformal CDC, Spyglass CDC, reset-domain crossings
- High-Speed Interfaces:
- PCIe Gen3 / Gen4 / Gen5, Ethernet (1G / 10G / 25G / 100G), DDR3 / 4 / 5, SerDes / GT transceivers, AXI / AMBA, Aurora, JESD204B / C
- DSP & Acceleration:
- FFT, FIR / IIR filters, polyphase, mixers, modems, image / video processing, ML inference, DSP slice mapping
- SoC Integration:
- Zynq UltraScale+ / Versal SoC, ARM Cortex-A / R5, RISC-V soft cores, AXI memory maps, Petalinux / U-Boot, baremetal bring-up
Education
Work Experience
- Own RTL design for Arista's 400GbE switch FPGA datapath across 4 board variants and 6 RTL releases, coordinating packet classification pipeline, MAC and PHY integration, and telemetry and statistics engines with influence across 40+ engineers across 5 IPTs.
- Partitioned the FPGA fabric across LUT, FF, BRAM, and DSP resources, hitting 78% LUT / 84% DSP / 71% BRAM utilization at 400 MHz pipeline clock with zero re-architecture spins across the program.
- Built UVM with SystemVerilog assertions testbenches with 240+ UVM sequences and constrained-random scenarios, hitting 99.4% functional coverage at sign-off across 3 RTL releases.
- Ran synthesis and implementation in Vivado with disciplined XDC / SDC constraints, taking WNS from -820 ps to +45 ps and shipping timing clean across all 27 PVT corners.
- Designed the clocking and reset architecture across 14 asynchronous clock domains with async FIFOs + 2-FF synchronizers and gray-code crossings, hitting zero CDC violations on Spyglass at sign-off.
- Integrated 100GbE + PCIe Gen4 stacks via vendor IP and custom AXI-Stream adapters, sustaining 400 Gbps line-rate throughput across 4 lanes and producing 3 custom RTL IPs adopted org-wide.
- Implemented hashing and CRC engines with hand-tuned pipelining and DSP-slice mapping, hitting 600 Gbps hashing throughput and consuming 38% fewer DSP slices than baseline.
- Integrated the FPGA fabric with the on-chip SoC on a Zynq UltraScale+ MPSoC, wiring up 14 AXI peripherals across 3 memory maps and partnering with embedded software on Linux-on-FPGA bring-up on first build.
- Built the team's CI/CD on Jenkins + TCL + Python, covering nightly synthesis + UVM regression on 38 RTL repos and cutting full regression-cycle time from 11 hours to 2.5 hours per regression cycle.
- Mentored 5 mid-career engineers through senior FPGA work, supporting 2 promotions to Senior, and authored the team's RTL code-style guide adopted across 4 product lines.
- Partnered with ASIC, Hardware, Embedded Software, and Verification teams across 4 SmartNIC product lines, authoring 24 spec and verification-plan docs that became the team's RTL baseline and onboarding 4 new FPGA engineers into the team's design-review and bring-up workflow.