Abhijit Kumar Sahu

HPC algorithm specialist with a proven record of designing and implementing high-performance parallel graph algorithms at billion-edge scale. Published at IEEE IPDPS (3 papers in 2026), Euro-Par, ICPP, HiPC (all CORE A venues). Achieved up to 900× speedup over state-of-the-art baselines; Euro-Par 2025 paper runners up for Best Paper Award.

Holds a US Patent on a high-concurrency in-memory database. Masters thesis explicitly recommended for the "Best MS Thesis" award. Deep expertise in CUDA kernel optimisation, PRAM algorithm design, out-of-core GPU computation, and dynamic graph processing.

Systems Engineer — Slingshot Interconnect (SDE-I) - Hewlett Packard Enterprise (HPE) - Bengaluru, India

(2025-02)

• Led end-to-end schema and models management within the HPE Slingshot mgmt-framework; designed reusable YANG model templates and successfully deployed the management framework on simulator container switches.
• Deployed separate REST management server packages for switch and control nodes, decoupling deployment units and improving modularity of the fabric configuration stack.
• Designed and deployed CRUD applications to interface seamlessly with the REST management servers, streamlining switch node configuration, telemetry monitoring, and state management for fabric administrators.
• Drove full Debian 13 (Trixie) migration of critical packages including (mgmt-common, libyang, yang-models); engineering workarounds for complex upstream dependency gaps to ensure cross-repository compatibility.
• Optimized CI/CD pipeline reliability by integrating comprehensive code coverage instrumentation across the management framework and eradicating flaky unit tests that previously blocked critical builds.
• Resolved multiple system critical bugs including fabric_bringup failures, UI access regressions along with multiple Debian package build failures and package split conflicts.

M.S. (Research) - Indian Institute of Technology Tirupati