Softblues

We need an AI Engineer who can work autonomously on complex agentic AI systems while contributing to our growing AI practice. You’ll work directly on client projects involving LLM agents, RAG systems, and AI model integration - not just building prototypes, but production systems that scale.
 

This isn’t a research role. This is hands-on engineering: building, deploying, iterating, and supporting AI applications that solve real business problems.
 

What You’ll Actually Do

Core Responsibilities (70% of time)

• Build agentic AI systems using LangChain, LangGraph, or similar frameworks
• Implement and optimize RAG pipelines with vector databases (PGVector, Pinecone, etc.)
• Integrate multiple LLM providers (OpenAI, Google Gemini, Anthropic Claude)
• Fine-tune models when off-the-shelf solutions don’t meet requirements
• Develop Python backends (FastAPI/Flask) that power AI applications
• Write production-quality code with tests, documentation, and proper error handling

Technical Ownership (20% of time)

• Design AI system architectures for new client projects
• Make build/buy decisions on AI tooling and infrastructure
• Conduct technical discovery with clients to understand requirements
• Estimate project complexity and technical feasibility

Team Collaboration (10% of time)

• Code reviews for junior AI engineers
• Knowledge sharing on AI patterns and best practices
• Client communication on technical approaches and trade-offs

Our Current Tech Stack 

AI/ML:

• LangChain, LangGraph for agentic systems
• OpenAI GPT-4, Google Gemini, Anthropic Claude
• PGVector for embeddings and retrieval
• Fine-tuning pipelines (OpenAI, Gemini)

Backend:

• Python 3.11+ (FastAPI primary, Flask acceptable)
• PostgreSQL with PGVector extension
• Google Cloud Run, Cloud Functions

Infrastructure:

• Google Cloud Platform (BigQuery, Cloud SQL, Cloud Run, Vertex AI)
• Firebase/Supabase for user management
• Docker for containerization

Frontend Integration:

• REST APIs consumed by React/TypeScript frontends
• Real-time capabilities (WebSockets, Server-Sent Events)

Development:

• Git/GitHub for version control
• Linear for project management
• Bolt.new for rapid prototyping (you won’t use this much)

Must-Have Skills
 

Technical Requirements

• 4+ years of professional Python development - You’ve built production systems, not just scripts. You understand design patterns, testing strategies, and code that scales.
• 2+ years building production AI/LLM applications - Real systems serving real users, not just experiments or prototypes. You’ve handled model deployment, monitoring, and iteration based on production feedback.
• Expert-level Python engineering - You write clean, maintainable, testable code. You can explain the difference between "staticmethod" and "classmethod" and actually care about it. You’ve debugged memory leaks, optimized performance bottlenecks, and know when to use async/await.
• Deep LangChain/LangGraph experience - You’ve built multi-agent systems with complex orchestration. You understand the framework’s internals well enough to work around its limitations. You’ve debugged agent loops, optimized token usage, and handled error states gracefully.
• RAG implementation expertise - You’ve built RAG systems that actually work in production. You understand chunking strategies, embedding models, hybrid search, re-ranking, and when RAG isn’t the right solution. You’ve tuned retrieval quality beyond the tutorial examples.
• Vector database expertise - PGVector, Pinecone, Weaviate, or equivalent. You’ve designed schemas, optimized queries, and handled millions of embeddings. You understand the trade-offs between different distance metrics.
• Production LLM integration - OpenAI, Anthropic, Google Gemini in real applications. You’ve handled rate limiting, cost optimization, prompt engineering, context management, and fallback strategies. You know how to debug hallucinations and improve consistency.
• Backend API development - FastAPI or Flask in production. You’ve designed RESTful APIs, handled authentication, managed database connections, implemented error handling, and written API documentation. You understand HTTP status codes and when to use which one.
• SQL and database design - PostgreSQL specifically. You write efficient queries, design normalized schemas, use indexes appropriately, and understand when to denormalize. You’ve debugged slow queries and optimized database performance.

Working Style Requirements

• Self-managed QA mindset - You write unit tests, integration tests, and think about edge cases before they become production bugs. You don’t need a QA engineer to tell you something doesn’t work.
• Pragmatic problem-solver - You prefer “simplest thing that works” over perfect solutions. You know when to refactor and when to ship. You understand technical debt is a tool, not a failure.
• Comfortable with ambiguity - Client requirements evolve mid-project. You adapt architecture without rebuilding from scratch. You ask clarifying questions but don’t wait for perfect specifications before starting.
• Strong technical communicator - You explain technical trade-offs to non-technical stakeholders. You write clear documentation. You give useful code review feedback. You can justify your architectural decisions.
• European timezone - Core hours overlap with UK (GMT/BST). You’re available for morning standups and client calls during UK business hours.

We are looking for a motivated Junior QA Engineer to join our dynamic team, focusing on web and AI-driven projects. This role is ideal for someone early in their career who is passionate about ensuring software development and has an interest in growing into a more advanced BA role in the future.

You will collaborate with developers, product managers, to test innovative web and AI applications, identify defects, and ensure an exceptional user experience.

Responsibilities

• Perform manual testing of web and AI applications to identify bugs, defects, and usability issues.
• Create, execute, and maintain test cases and test scenarios based on project requirements.
• Document and report defects with clear steps to reproduce, using bug-tracking tools (e.g., Jira).
• Collaborate with the development team to verify fixes and ensure issues are resolved.
• Participate in requirement analysis and provide feedback to improve product quality.
• Assist in regression, functional, and exploratory testing activities, including testing AI model outputs and web interfaces.

• Contribute to the continuous improvement of QA processes and methodologies, with an eye toward future automation opportunities.

   

Requirements

• +1 year of experience in software testing or a related field (internships or personal projects are a plus).
• Basic understanding of software testing principles, methodologies, and the software development lifecycle (SDLC).
• Familiarity with bug-tracking tools (e.g., Jira, Trello) and test management tools (e.g., TestRail).
• Interest in web and AI technologies, with exposure to testing web applications (e.g., REST APIs, browser-based interfaces) and AI-driven features (e.g., chatbots, recommendation systems).
• Knowledge of relevant tech stack: JavaScript, Python (for testing AI components), Postman (for API testing), and browser developer tools (e.g., Chrome DevTools).
• Strong attention to detail and problem-solving skills, with a proactive mindset for learning and growth in the QA field.
• Good communication skills to collaborate effectively with cross-functional teams working on web and AI solutions.
• Upper-intermediate English proficiency (written and verbal).
• (Optional) Basic knowledge of automation tools (e.g., Selenium, Cypress) or scripting languages (e.g., Python, JavaScript) is a plus.
• A strong desire to develop and advance in the QA direction, with potential interest in automation, performance testing, or AI-specific testing in the future.

We need an AI Voice Engineer who can work autonomously on complex agentic AI systems while contributing to our growing AI practice. You’ll work directly on client projects involving LLM agents, RAG systems, and AI model integration - not just building prototypes, but production systems that scale.

This isn’t a research role. This is hands-on engineering: building, deploying, iterating, and supporting AI applications that solve real business problems.

• LiveKit Agents: Open-source, vendor-agnostic, built-in telephony (SIP)
• Google ADK: Gemini-native, Google Cloud integration, Vertex AI
• Realtime/Live APIs: OpenAI Realtime API, Gemini Live API
• STT: Deepgram, AssemblyAI, Google STT, Whisper (OpenAI)
• TTS: ElevenLabs, Cartesia, Google TTS, Rime, PlayHT
• LLM/Reasoning: Gemini 2.5 Flash / Pro, GPT-4o / GPT-4.1, Claud
• VAD & Turn detection: Silero VAD, Semantic Turn Detection, Endpointing configuration
• Testing & Evaluation: Hamming AI, Coval, Braintrust, Langfuse, Custom LLM-as-Judge
• Supporting technologies: WebSocket, WebRTC, Python (primary), Node.js, GCP, AWS, or self-hosted
• Telephony (nice to have): SIP, PSTN, Twilio Voice

A senior AI engineer who builds AI applications on Google Cloud using Gemini models. They should have deployed real projects to production (not just experiments), understand Google's AI tools, and be able to work with enterprise clients.

• 5+ years Python software engineering overall
• 2+ years working with AI/ML
• Experience building and deploying AI apps with Gemini on Google Cloud: 1–2 production projects with specific examples (e.g., chatbots or automations); experiments without deployment are insufficient.
• Familiarity with Google ADK (Agent Development Kit): Hands-on building or studying agents, including multi-agent systems or Google Cloud integrations.
• Work with Vertex AI Search and RAG systems on Google Cloud: Built search or Retrieval-Augmented Generation solutions, describable in detail (e.g., vector DB integration, semantic search optimization).
• Recent Google Cloud AI project: Latest work using Gemini 2.5 (or similar) within the last 6–12 months;

Description:

General InformationWe are seeking 1-2 experienced Cyber Security Experts to support investor and risk management teams operating in the UK, US, Japan, and South Korea. The ideal candidate will be based in Japan or South Korea, fluent in Japanese or Korean, and capable of working directly with senior-level clients and internal teams.This is a full-time position, but we are also open to part-time or hourly engagements. Candidates must be available to work with at least 4–6 hours of overlap with Japan or Korea business hours (preferably full coverage).

You’ll be joining a global cyber-security delivery team, collaborating closely with internal stakeholders and external clients.

Key Responsibilities

• Conduct cyber maturity assessments using NIST Cybersecurity Framework.
• Use internal cyber questionnaire tools to evaluate client security posture.Support identity and access management (IAM) initiatives.
• Work with senior-level clients across investor and risk management teams.
• Contribute to internal reporting systems such as timesheets, HR portals, etc.
• Collaborate with local junior cyber resources in Japan and Korea; provide senior-level guidance.
• Deliver hands-on, broad-spectrum cyber expertise ("Jack of all trades" mindset).
• Interface with internal delivery teams and provide updates to management.

Requirements

• 5+ years in Cyber Security, preferably with experience in financial services or risk management.
• Strong working knowledge of the NIST Cybersecurity Framework.
• Proven experience with identity and access management (IAM).
• Comfortable engaging with executive and senior client stakeholders.
• Ability to work independently and remotely.
• Prior experience with cybersecurity assessments, risk tools, and evaluation processes.
• Availability for 4-8 hours overlap with Japan/Korea time zones.
• Fluent in Japanese or Korean (local candidates strongly preferred).

Detailed Required Skills:

• IT Security
• Enterprise Cybersecurity
• Customer Identity and Access Management (CIAM)
• Japanese
• NIST
• Cybersecurity Operations
• Cybersecurity Maturity Model Certification (CMMC)