Akshay Borse — Senior Software Engineer · AI infrastructure, distributed systems, developer platforms

LinkedIn · Featured Projects · Research · Email

Senior Software Engineer with 8+ years at Microsoft and Amazon building AI infrastructure, distributed systems, security platforms, and developer tooling — including AI security for 237K+ agents across 13K+ organizations and event orchestration at 100K+ TPS.

Selected accomplishments from professional roles — not open-source repository statistics.

• AI-agent governance and risk enforcement across 13K+ enterprise tenants
• Platform standardization across ~50 services and 40 engineers, saving 40–50 engineering-weeks
• Deployment across commercial and US government cloud environments (GCC, GCCH, DoD)

AI security and governance for agentic identities in Microsoft Purview Insider Risk Management.

• Risk policy and enforcement for 237K+ active AI actors across 13K+ enterprise tenants
• Audit-history enrichment across 40+ global AD forests and three sovereign cloud regions
• Multi-region ingestion across 28 Azure regions processing 135K+ daily risk signals with compute isolation

Tech: C# · .NET · Azure Functions · Cosmos DB · Event Hubs · Kusto · Bicep · ARM · Microsoft Purview · Geneva

Platform engineering, distributed systems, production ML, and observability at scale.

• DSL-driven microservice bootstrap framework — weeks to hours setup, 40–50 engineering-weeks saved across ~50 services
• Event orchestration at 100K+ TPS across 16 microservices and 8 business-critical workflows
• SageMaker and Bedrock ML systems — $28M annualized delivery impact, 7% sales conversion lift, 13–16% fulfillment gains
• CloudWatch logging optimization — $250K+/month cost reduction; 60% fewer Sev-2 alerts

Tech: Java · Python · gRPC · REST · Docker · ECS · Lambda · DynamoDB · S3 · Aurora · Step Functions · App Mesh · Redis · Kinesis · CloudWatch · OpenSearch · SageMaker · Bedrock

LeetDesign Interactive system-design simulator for modeling scalability, fault tolerance, and architectural trade-offs through executable service and workload models. Concepts: capacity modeling · failure domains · load shaping · partition simulation	FinanceOS AI-powered financial intelligence platform that transforms transactions into explainable insights and personalized recommendations through deterministic processing and LLM-assisted reasoning. Concepts: deterministic pipelines · explainable AI · guarded LLM reasoning · multi-currency ledgers Live demo
Arsenal Developer-tooling platform for project scaffolding, workflow automation, repository analysis, validation, and context-efficient AI coding workflows. Concepts: scaffolding · repo analysis · workflow automation · AI context optimization	Orbit Offline-first productivity application for managing tasks, recurring routines, habits, workload planning, and personal analytics. Concepts: offline-first sync · local persistence · habit loops · workload analytics

Ad-Infinitum — F2FS Garbage-Collection Research

Linux filesystem research exploring F2FS garbage-collection policies through kernel instrumentation, performance benchmarking, and latency optimization. Published survey comparing NILFS and F2FS cleaning policies with compile-bench throughput analysis.

Concepts: F2FS · log-structured filesystems · garbage collection · kernel instrumentation · latency benchmarking

Published paper · Repository

Interactive knowledge-graph platform for exploring cognitive patterns, behaviors, decisions, and feedback loops.

Repository · Live demo

Java · C# · Python · JavaScript · SQL · C/C++ (systems & filesystem research)

Agentic AI governance · RAG · prompt engineering · Amazon SageMaker · Amazon Bedrock · Anthropic LLMs · production ML serving · quantile regression · XGBoost classification

Microservices · REST · gRPC · event-driven architecture · high-throughput orchestration · fault tolerance · horizontal scaling · Redis

Azure: Functions · Cosmos DB · Event Hubs · Kusto · Bicep · ARM · sovereign cloud (GCC, GCCH, DoD) · AWS: ECS · Lambda · DynamoDB · S3 · Aurora · Step Functions · App Mesh · Kinesis · Docker

Cosmos DB · DynamoDB · Aurora · S3 · event replay · vector search · multi-tenant data models

CloudWatch · Geneva · OpenSearch · Kibana · Kusto · telemetry standardization · dynamic log-level APIs · MTTR reduction

Unit & integration testing · load & stress testing · canary analysis · fault injection · automated rollback · CI/CD · infrastructure as code · TRA compliance

AI identity risk scoring · tenant-configurable policy enforcement · insider risk management · OAuth 2.0 · sovereign cloud compliance

flowchart TB
  AI["AI Infrastructure"] --> SEC["AI Security & Governance"]
  AI --> ML["Production ML"]
  SEC --> DS["Distributed Systems"]
  ML --> DS
  DS --> CLOUD["Cloud Platforms"]
  CLOUD --> DEV["Developer Tooling"]
  DS --> DEV
  DEV --> DEL["Platform Delivery"]
  CLOUD --> DEL

Interested in AI infrastructure, distributed systems, or developer platforms?

LinkedIn · Portfolio · Email · GitHub

Redmond, WA