Principal AI Engineer (AI Evangelist)

Job Title: Principal AI Engineer (AI Evangelist)
Location: New York City NY
Setting: Hybrid Role (4 days onsite - 1 day remote)
Type: FTE role with Mphasis

AI Evangelist -Hands-On:
A hands-on AI Evangelist in a financial organization plays a vital role bridging the gap between cutting-edge AI technologies and practical business needs. This is a senior role not only involves technical development but also demands effective communication and advocacy to facilitate the responsible adoption of AI in finance.

Key Responsibilities as Evangelist

• Build and demonstrate AI-powered solutions for financial applications preferably in investment banking Trading or insurance environments.
• Translate complex AI concepts into actionable business value for both technical and non-technical stakeholders simplifying information for internal teams and executive leadership.
• Lead workshops seminars and training sessions for teams across the organization promoting AI literacy and upskilling staff in banking investment or insurance environments.
• Ability and /or experience in authoring technical blogs white papers and internal documentation that explain the impact and possibilities of AI in the financial domain.
• Experience on working on advisory capacity to CxO Head of Engineering Head of Architecture on technical strategies
• Act as a visible presence at industry conferences webinars and external forums to position the organization as a leader in responsible AI use within finance.
• Partner with compliance risk and IT teams to ensure all AI solutions meet strict regulatory and ethical standards prevalent in financial services.
• Prototype test and deploy AI models that address market forecasting customer insights automated underwriting or anti-money laundering strategies.

Key Technical and Design Responsibilities

• Build deploy and manage both agentic AI architectures and generative code systems ensuring scalable and secure integration of technologies like LLMs code generators and automation agents within production workflows.
• Oversee technical design implementation and code reviews-especially for code created or assisted by AI tools-maintaining high standards of security performance and maintainability in Python and other programming languages.
• Develop robust testing and validation protocols for AI-generated code and agent behavior including prompt engineering debugging and post-deployment monitoring for unusual failure patterns or compliance issues.
• Lead technical teams mentor junior engineers and set excellence in software engineering practices; create documentation and establish guidelines for both human and AI-driven contributions
• Collaborate with cross-functional stakeholders (DevOps security product and business leaders) to ensure rapid safe adoption of agentic and generative AI features

Essential Qualifications

• Bachelors or masters degree in computer science Data Science Finance or related field.
• Experience in one or many of the high-level programming languages like C Java C#
• Good understanding of Typescript and other JS framework for UI development
• Strong hands-on experience with Python SQL and AI/ML frameworks (e.g. TensorFlow PyTorch) as applied to financial data and workflows
• At least 4 years working in AI roles within finance fintech or technical consulting preferably with exposure to regulatory environments.
• Deep knowledge of AI ethics compliance Guardrails data privacy and compliance trends relevant to the financial sector.
• Excellent communication stakeholder engagement and technical storytelling abilities.73
• Demonstrated ability to manage multiple priorities and projects while maintaining strategic alignment and rigorous attention to detail.

Desired Traits

• Passion for driving innovation and adoption of AI in highly regulated settings.
• Effective at stakeholder education-from boardroom to engineering teams.
• Strong problem-solving mindset able to translate technical outputs into practical business recommendations.
• Integrity and diligence with a commitment to both organizational objectives and ethical AI deployment.

Essential Experience & Skills

• Advanced Python expertise plus experience with other major backend languages (e.g. Java C Go) and modern AI/ML toolkits
• Demonstrated proficiency in designing validating and launching code-generation systems and agentic workflows strong familiarity with prompt engineering and AI model deployment
• Track record of hands-on technical leadership within agile teams overseeing both human and AI-generated codebases and ensuring auditability explainability and compliance at scale
• Expertise in code review automated testing and documentation standards for mixed human/AI development environments

Desired Traits

• Analytical mindset with a passion for innovation experimentation and best practices in AI-enhanced software engineering
• Effective communicator comfortable translating complex AI behaviors and code-gen strategies for both technical and business audiences
• Champion of ethical AI development security consciousness and responsible agent operation in critical production settings

Essential AI Design and Architecture Skills

• Prompt Engineering: Crafting structured prompts to drive deterministic and reproducible outputs from LLMs using techniques like chain-of-thought and few-shot prompting
• Context Engineering: Dynamically injecting relevant external data into prompts; designing and managing context windows handling retrieval noise and context collapse in long-context
• Fine-Tuning & Model Adaptation: Using methods like LoRA/QLoRA for domain adaptation managing data curation pipelines and monitoring overfitting versus generalization-especially in high-stakes environments
• Retrieval-Augmented Generation (RAG): Building LLM workflows with external knowledge integration engineering embeddings and retrieval pipelines for high recall and precision
• Agentic Design: Orchestrating LLM-driven agents capable of multi-step reasoning tool use and autonomous state management-including fallback strategies for error
• Production Deployment: Packaging models and agentic systems as scalable APIs with robust pipelines for latency concurrency and failure isolation including container orchestration or serverless deployment
• LLM Optimization: Applying quantization pruning and distillation to optimize performance and cost; benchmarking for speed accuracy and hardware utilization
• Observability & Monitoring: Implementing logging tracing dashboards and alignment monitoring for prompts responses and agent behaviors
• Core SDLC AI Integration: Using generative AI for requirement refinement technical design blueprinting architecture review API and schema auto-generation and cross-functional artifact production
• Security & Compliance: Building guardrails to enforce data privacy compliance with regulations and responsible use of LLMs particularly in sensitive or regulated environments.
• Modern Deep Learning: Mastery of frameworks including TensorFlow PyTorch and HuggingFace Transformers with proven expertise in transformers CNNs RNNs and attention mechanisms for custom and state-of-the-art model

Essential AI Tools

• GitHub Copilot: Mainstream AI-powered code generation and completion for major languages widely integrated into enterprise SDLC
• ChatGPT/GPT-4/Vision: Prompt-driven code assistance architecture brainstorming documentation generation and natural language requirement mapping
• SonarQube: AI-powered static code analysis and vulnerability detection for code security and quality assurance across 1
• Jira (with AI plugins): AI-enhanced project management backlog refinement and sprint planning-crucial for orchestrating product delivery at scale
• Claude Code: Multi-step code generation and agentic orchestration especially suitable for agent-based SDLC
• Datadog and Dynatrace: Proactive AI in monitoring predictive analytics and incident response for production reliability and observability.
• RAG frameworks like Langchain Langraph LlamaIndex Graph RAG
• Graph database -RD4j Neo4j and timeseries database
• Embeddings & Vector Databases: Understanding embeddings vector search vector DB platforms (FAISS Pinecone Chroma Weaviate) and semantic retrieval
• Observability & Evaluation: Setting up logging debugging and automated quality evaluation for RAG applications (e.g. with TruLens Streamlit dashboards).
• Containerization/DevOps: Packaging with Docker or similar using cloud/AWS/Azure integrations for scalable deployments.

Nice to Have AI Tools

• Sourcegraph Cody: LLM-powered search code context awareness and auto-completion over vast enterprise repositories
• Cursor/Codex/Windsurf: AI-native development workflow management; designed for large-scale coding and agentic workflows
• Amazon Q (AWS): AI-driven code architecture recommendation and AWS-native code and cloud resource management
• Synapt SDLC Squad: Multi-agent generative AI platform for end-to-end SDLC automation code review and compliance.
• Bitbucket/GitLabs (with AI modules): AI-assistance in code review merge requests release management and security
• Figma (AI for Design/Prototyping): AI documentation prototyping and developer handoff for frontend/UI

Nice to Have architecture Skills

• Lightweight Architecture ADLs: Using architecture definition languages (ADLs) to leverage LLMs for generating structural constraints fitness functions and architecture governance
• Automated UI/UX Prototyping: Leveraging generative tools (e.g. Claude RunwayML) for rapid wireframe and design generation from requirements or stakeholder
• End-to-End SDLC Automation: Experience with multi-agent platforms and orchestration tools for automating requirement gathering code review compliance and release 1
• Cost-Efficiency & Sustainability: Implementing sustainable AI practices carbon-aware scheduling and model lifecycle management for production
• Emergent LLM Platforms: Experience integrating and orchestrating new LLMs open-source agents vector DBs and hybrid architectures beyond mainstream offerings
• Ethical AI & Governance: Leading the definition of internal standards and policies for responsible bias-safe LLM operation and agentic workflows
• Domain-Specific Knowledge: Deep knowledge of applying LLMs and generative AI in specialized contexts such as finance Banking or other regulated domains.