Software Engineering Intern
Deepshard, Inc. · United States
- Hybrid
- Full-time
- $72,000 / year
- United States
Job highlights
- Build AI agent integrations for data sovereignty.
- Reverse-engineer APIs and build tool servers.
- Work on Slack, Google Workspace, Spotify, Instacart.
- Gain experience with agentic AI systems.
- High ownership, fast-paced environment.
About the role
Software Engineering Intern
Location: Remote
Compensation: $6,000 USD/month stipend (1st month)
About Deepshard
Deepshard is building Truffle¹ — an always-on agentic workstation with an AI agent running entirely on-device with zero cloud dependency. We give individuals and organizations true data sovereignty and IP protection. No data leaks to Big Tech, no surveillance risk. Truffle¹ processes everything locally, builds deep personal/organizational memory, and operates as an always-on, private intelligence system.
The Role
We're looking for a Software Engineer Intern to build out integrations that let Truffle take actions and ingest information. Your job is to make the agent useful by building tools — whether that means reverse-engineering undocumented APIs, building MCP tool servers, or wiring up integrations with platforms like Slack, Google Workspace, Spotify, Instacart, and dozens more. You'll work at the intersection of AI agent systems and platform engineering. The agents are smart, but they need well-built tools to act on a user's behalf. You'll build those tools.
What You'll Do
- Reverse-engineer APIs — Discover and document undocumented API surfaces that consumer platforms actually use. Handle authentication flows, anti-bot measures, session management, and rate limiting.
- Build MCP tool servers — Design and implement tool interfaces that AI agents can call reliably. Define clean input/output schemas, handle edge cases gracefully, and make tools that an LLM can reason about effectively.
- Ship integrations end-to-end — From API discovery through client library to production tool server. Platforms will include messaging (Slack, Discord), productivity (Google Workspace, Notion), commerce (Instacart, Amazon), media (Spotify, YouTube), and others as the product evolves.
- Work within agentic systems — Understand how LLM agents select and invoke tools, how context flows through multi-step workflows, and how to design tool APIs that minimize agent errors.
- Handle the messy parts — Parse unstructured data, manage flaky external services, build retry/backoff logic, persist state across runs, and write honest documentation about what works and what doesn't.
What We're Looking For
- 1–3 years of professional software engineering experience (Python preferred)
- Demonstrated ability to reverse-engineer web APIs using browser devtools, traffic interception, or similar techniques
- Experience building or consuming HTTP APIs with proper error handling, pagination, and auth management
- Familiarity with agentic AI systems — you understand how tool-use works with LLMs (MCP, function calling, or similar frameworks)
- Strong debugging instincts and a bias toward shipping working software over polished prototypes
Nice to Have
- Experience with the Model Context Protocol (MCP) or similar tool-serving frameworks
- Background in web scraping, data extraction, or browser automation
- Familiarity with running local/open-source LLMs (e.g., Qwen, Llama, Mistral)
- Experience with background workers, scheduled jobs, and state management patterns
- Understanding of anti-bot/anti-scraping countermeasures and how to work around them responsibly
- Hardware knowledge for potential physical integrations (controlling light switches, home monitoring system etc)
Hiring Process
- Initial interview — Technical conversation about your experience with APIs and integrations
- Take-home challenge — A focused project that mirrors real work you'd do on the team (reverse-engineering APIs, building an MCP tool server, connecting an AI agent to real platforms)
- Final interview — Deep-dive on your submission and team fit
- Paid trial (1 month) — Work with the team on real projects. If it's a good fit on both sides, you stay on.
Why Truffle
You'll be building something genuinely different — a private, on-device AI system that respects user sovereignty. The integration layer you build is what makes the agent actually useful in people's daily lives. Small team, high ownership, real impact on the product from day one. If a fast paced, high stakes environment where you learn new things at an expedited pace excites you, you will like working with us.
Key skills/competency
- Software Engineering Intern
- API Reverse Engineering
- Tool Development
- Platform Integrations
- AI Agent Systems
- HTTP APIs
- Python
- Debugging
- Web Scraping
- LLM Tool Use
Skills & topics
- Software Engineering Intern
- Python
- API Reverse Engineering
- AI Agents
- Platform Engineering
- Integrations
- HTTP APIs
- Tool Development
- Remote
- Internship
How to get hired
- Tailor your resume: Highlight Python, API reverse-engineering, and AI agent experience.
- Showcase projects: Detail personal or academic projects involving API integrations or tool development.
- Prepare for technical interviews: Brush up on HTTP APIs, authentication, and debugging techniques.
- Understand Truffle's mission: Emphasize your interest in data sovereignty and on-device AI.
- Engage with the challenge: Approach the take-home assignment methodically, demonstrating problem-solving skills.
Technical preparation
Practice reverse-engineering web APIs.,Build and test HTTP APIs with error handling.,Understand LLM function calling concepts.,Write Python code for API integrations.
Behavioral questions
Describe a complex technical problem you solved.,How do you handle ambiguous requirements?,Tell me about a time you shipped code quickly.,How do you document your work?,How do you learn new technologies rapidly?
Frequently asked questions
- What technical skills are most important for the Software Engineering Intern role at Deepshard?
- For the Software Engineering Intern position at Deepshard, strong skills in Python, API reverse-engineering (using tools like browser devtools and traffic interception), and experience with building or consuming HTTP APIs are crucial. Familiarity with AI agent systems and how they interact with tools (like MCP or function calling) is also highly valued.
- How does Deepshard ensure data sovereignty with Truffle?
- Deepshard ensures data sovereignty by building Truffle as an AI agent that runs entirely on-device, with zero cloud dependency. This means all data processing and memory building occur locally, preventing data leaks to third-party services and eliminating surveillance risks associated with cloud-based AI.
- What is the hiring process for the Software Engineering Intern at Deepshard?
- The hiring process begins with an initial technical interview focusing on API and integration experience. This is followed by a take-home challenge that simulates real work. Successful candidates then proceed to a final interview for team fit, and potentially a paid one-month trial period to work on live projects.
- Can I apply for the Software Engineering Intern role if I have less than 1 year of professional experience?
- The ideal candidate for the Software Engineering Intern role has 1-3 years of professional software engineering experience, with Python preferred. However, demonstrated ability in reverse-engineering APIs and building HTTP APIs, even through academic projects or personal work, can be a strong differentiator. Focus on showcasing relevant skills and projects.
- What kind of integrations will a Software Engineering Intern build at Deepshard?
- As a Software Engineering Intern, you will build integrations to enable the Truffle AI agent to take actions and ingest information. This involves reverse-engineering APIs, building tool servers (MCP), and creating end-to-end integrations with platforms such as Slack, Google Workspace, Spotify, and Instacart.
- What is the 'Model Context Protocol' (MCP) mentioned in the job description for Deepshard?
- The Model Context Protocol (MCP) is a framework that Deepshard uses to design and implement tool interfaces for AI agents. As a Software Engineering Intern, you would be responsible for building tool servers that adhere to MCP specifications, ensuring AI agents can call these tools reliably and reason about their input/output schemas.