Arun
Bongu

Full-stack Node.js/Express web app for planning and executing DR tests across AWS, Azure, GCP, and Oracle Cloud. Validates provider credentials from environment variables (AWS STS, Azure AD, GCP SA, OCI). Ships with 3 agent instruction files: SKILL.md (agent capabilities), agent.md (EC2 DR planner), instructions.md (system prompt for DR plan generation). Deployed serverless on Vercel via serverless-http.

// Validate credentials from env or named profile
if (!effectiveAccountId && (providerProfile || hasEnvCredentials(provider))) {
  const validation = await validateProviderProfile(provider, providerProfile, primaryRegion);
  effectiveAccountId = validation.accountId;
}
// Build DR steps: within-region AZ or cross-region failover
result.steps.push({ title: `Execute chaos tool test: ${chaosTool}`,
  detail: `Trigger ${chaosTool} against ${provider} ${serviceLabel}.` });
result.steps.push({ title: 'Validate recovery',
  detail: 'Confirm health, traffic routing, and data integrity after failover.' });

Production Next.js 16 / React 19 / TypeScript dashboard for launching, monitoring, and terminating AWS EC2 instances. Auto-resolves Amazon Linux 2023 AMI via SSM Parameter Store with 8-region static fallback. Includes SSH copy-to-clipboard, developer console log drawer, login page. Ships with SKILL.md (ops) and AGENTS.md (Next.js breaking-changes guard).

// Resolve latest Amazon Linux 2023 AMI via SSM Parameter Store
async function getLatestAMI(region: string): Promise<string> {
  const cmd = new GetParameterCommand({
    Name: "/aws/service/ami-amazon-linux-latest/al2023-ami-kernel-default-x86_64"
  });
  const res = await ssmClient.send(cmd);
  return res.Parameter?.Value ?? fallbackAMIs[region]; // 8-region static map
}

Python library for deploying AI/ML models to Amazon SageMaker using your own Docker container image. Supports both real-time inference endpoints and batch transform jobs. Published on PyPI as pip install inference-server. Full documentation on ReadTheDocs. Open-sourced by J.P. Morgan Chase & Co. under Apache 2.0.

Suite of Python algorithmic trading bots using the Interactive Brokers TWS API via ib_insync. RiskyOptionsBot buys 2-DTE SPY calls on 3 consecutive 5-min higher closes; thetagang runs The Wheel strategy (CSP → assignment → CC) for premium income; iron-condor executes a market-neutral spread strategy on QuantConnect.

if df.close.iloc[-1] > df.close.iloc[-2] and df.close.iloc[-2] > df.close.iloc[-3]:
    # 3 consecutive higher closes → enter 2-DTE SPY call, $5 OTM
    for optionschain in self.chains:
        for strike in optionschain.strikes:
            if strike > df.close.iloc[-1] + 5:
                self.options_contract = Option(
                    self.underlying.symbol, optionschain.expirations[1],
                    strike, 'C', 'SMART', tradingClass=self.underlying.symbol)
                trade = self.ib.placeOrder(self.options_contract, MarketOrder('BUY', 1))

Cloud infrastructure and FinOps portfolio: AWS Lightswitch Pattern (Lambda + EventBridge for scheduled cost control of non-prod environments), Terraform modules for AWS and GCP, GKE RBAC user personas + in-cluster API service accounts, Azure ML datastores/AutoML pipelines, and authored AZ-104 supplemental lab materials.