Job Specifications
AI Software Developer
About TechTorch
TechTorch is building the future of intelligent work. We help clients design, build, and deploy AI agents that automate complex, real-world workflows - delivering both reliability and measurable ROI. As part of our team, you won’t just be experimenting with ideas; you’ll be shipping production systems that create real impact across multiple industries. You’ll join a hands-on, fast-moving group that thrives on solving hard problems, pushing the boundaries of what AI can do, and turning breakthrough technology into practical results.
Role overview
We’re looking for an AI Software Developer to design, implement, and maintain agentic systems for our clients. You’ll work across LLMs, orchestration frameworks, and data pipelines to deliver robust, observable, and secure automations.
This is not a research-only role, and you won’t be stuck in endless proof-of-concepts. Every project you take on will target production and client systems, with real use cases and measurable impact.
You’ll be joining a hands-on team that thrives on solving challenging problems, shipping production-grade systems, and pushing the boundaries of what AI can do in real business environments.
What you’ll do
Design and implement agent architectures and multi-step orchestrations in Python
Build with frameworks and runtimes such as LangGraph, LangChain, n8n, AWS Bedrock, and similar tools
Integrate and evaluate frontier models, e.g., Google Gemini, OpenAI GPT family, and Anthropic Claude
Develop connectors to SaaS and enterprise systems via APIs, webhooks, queues, and events
Implement tools, memory, retrieval, and planning strategies for reliable agent behavior
Author a reference agent template / internal SDK (LangGraph patterns, tool adapters, memory/plan primitives) adopted across agents
Productionize prototypes: CI/CD, packaging, containerization, and deployment to cloud
Instrument agents for observability: logging, tracing, evaluation, offline analysis, and guardrails
Collaborate with client teams to capture requirements, run experiments, and iterate quickly
Facilitate design thinking workshops to understand business requirements, map user journeys, and translate needs into agent capabilities and success metrics
Apply security, privacy, and safety best practices for data handling and model usage
Minimum qualifications
Hands-on experience building LLM-powered agents or complex automations
Strong Python skills, including async programming, testing, and packaging
Experience with TypeScript (in addition to Python) is a plus, as it’s an emerging language in our stack
Practical experience with LangGraph and/or LangChain, plus one or more of: n8n, AWS Bedrock, Airflow, Prefect, Temporal, or similar orchestration tools
Working knowledge of core model ecosystems: Gemini, OpenAI, and Claude (prompting, tool use, structured output, function calling)
Solid data foundations: vector stores, embeddings, RAG patterns, chunking, indexing, and data quality
Software engineering fundamentals: Git, code reviews, CI, containerization, cloud services (AWS preferred)
Familiarity with ADK (Agent Development Kit) or equivalent frameworks for building and standardizing agent behaviors
Nice to have
Experience with agents in production: evaluation frameworks, red-teaming, safety filters, rate-limit and cost controls
Retrieval and data management at scale: document processing, ETL, and governance
AI orchestration with frameworks and runtimes such as LangGraph, LangChain, n8n, AWS Bedrock, Azure Foundry, and similar tools
Backend development: FastAPI, Flask, Node.js, or Go; relational and NoSQL databases (e.g., Postgres, MongoDB) and modern platforms like Supabase
Workflow automation with third-party apps and enterprise systems (Slack, Google Workspace, Salesforce, ServiceNow, etc.)
Observability stacks: OpenTelemetry, Langfuse, Weights & Biases, or custom analytics
How you work
Systems thinker who can break down ambiguous client problems into shippable agent capabilities
Bias to action, comfort with rapid experimentation and clear measurement
Strong written communication and partner‑facing collaboration skills
Location and working model
Remote‑first with occasional client or team onsite sessions as needed
Benefits and compensation
Competitive salary
Flexible time off and remote work support
Professional development budget for conferences, certifications, and experimentation
If you’re excited to build dependable AI agents that solve real problems, we’d love to talk.