Akarsh Jha

• Deployed and integrated open-source model on DGX GPU servers for enterprise speech-to-text.
• Built and administered an on-premises development server with secure SSH access.
• Implemented self-hosted CI/CD pipelines using GitLab Runner for automated deployments.
• Enabled secure remote access to internal AI infrastructure using Tailscale.
• Automated infrastructure and deployment workflows to improve reliability and operational efficiency.
• Managed user provisioning on API credentials, AI model assignments, and usage budgets for self-hosted LLM infrastructure.

• Fine-tuned open-source LLMs on domain-specific datasets to improve the performance of AI agents.
• Developed automated data preprocessing pipelines to enhance dataset quality and streamline workflows.
• Benchmarked AI models using F1-score and other performance metrics for model optimization.
• Optimized production model on NVIDIA DGX GPU infrastructure to improve latency and inference efficiency.
• Managed cloud-based training and experimentation workflows on AWS to accelerate AI model development.
• Developed an agentic CLI featuring parallel AI agents to automate specialized development and operational workflows.

• Built and shipped privacy-first consumer software from concept to production.
• Designed scalable backend systems, APIs, and application architecture across multiple products.
• Developed cross-platform applications, browser extensions, and AI-powered features.
• Led product design, user experience, branding, and technical strategy.
• Managed product releases, user feedback, and continuous feature improvements.
• Grew products organically through iterative development and community-driven marketing.