Founding AI Engineer (Agentic AI)

4y relevant experience

Qualified

Executive Summary

The candidate is a senior backend engineer with 7 years of experience and a strong recent pivot into production agentic AI systems, making them one of the more technically aligned candidates for this Founding AI Engineer role. Their current work building multi-service LangGraph pipelines, RAG architectures, and agentic ecosystems on AWS directly mirrors the technical stack and responsibilities described by AlpacaRelay. Key concerns include the absence of public code artifacts for independent validation, missing experience with several preferred tools (LangSmith, LangFuse, CrewAI, LlamaIndex, MCP servers), and limited signal on ML fundamentals depth beyond applied LLM integration. Their India-based location and B2B contract model should be clarified to ensure alignment with the role's remote/B2B structure. Overall, the candidate represents a solid FIT candidate who warrants a structured technical interview to validate code quality and ML depth before proceeding to an offer.

Top Strengths

✓Production-grade LangGraph agentic pipeline experience with real users and multi-tenant architecture
✓Strong full-stack backend engineering foundation across Python, Go, Node.js, and TypeScript enabling cross-stack ownership
✓Proven cloud deployment expertise (AWS AgentCore, ECS, Terraform, Docker, Kubernetes) critical for a founding engineer role
✓Hands-on RAG implementation experience including hybrid search, reranking, semantic caching, and streaming — directly applicable to content generation products
✓7 years of backend experience with demonstrated progression from developer to senior engineer, suggesting adaptability and learning velocity

Key Concerns

!Absence of public code artifacts (GitHub, open source) makes independent technical validation difficult and raises questions about transparency and community engagement expected of a founding engineer
!Missing experience with several key required tools (LangSmith, LangFuse, CrewAI, LlamaIndex, MCP servers, SciPy) and no PhD/advanced degree — may require significant onboarding ramp for the full expected tech stack

Culture Fit

65%

Growth Potential

High

Salary Estimate

$80,000 - $100,000 (India-based, B2B contract — likely competitive in range given 7 years experience but geography may influence expectations)

Assessment Reasoning

The candidate is assessed as FIT with a score of 72, driven primarily by their directly relevant and current production experience building LangGraph-based agentic pipelines, multi-tenant RAG systems, and cloud-deployed AI applications — all of which are core to this Founding AI Engineer role. They meets the 2+ year minimum requirement substantially (7 years total, ~4 years AI-relevant), demonstrates Python strength, and has real exposure to agentic frameworks, vector search, and cloud infrastructure. The score is held back from the 80s by the absence of code samples or GitHub artifacts (limiting quality verification), missing experience with several required tools (LangSmith, LangFuse, CrewAI, LlamaIndex, MCP servers), no open-source contributions, and limited explicit ML fundamentals depth. These gaps are meaningful for a founding engineer role but are not disqualifying given the strength of their production AI engineering background. A technical interview with a coding challenge and architecture deep-dive is strongly recommended to confirm fit before advancing to offer stage.

Interview Focus Areas

Deep dive into the Cognitosoft LangGraph agentic pipeline — architecture decisions, trade-offs, failure modes, and lessons learnedAssessment of startup ownership mindset: ask about times they drove technical decisions independently, shipped under ambiguity, or wore multiple hatsEvaluate ML fundamentals depth: NumPy, SciPy, embedding models, evaluation metrics, and understanding of model internals beyond API usageProbe knowledge of observability and evaluation: how they monitor AI outputs, detect drift, and measure RAG quality in productionTechnical challenge or take-home to assess code quality, Python craftsmanship, and ability to rapidly prototype an agentic workflow

Code Review

FairSenior Level

No code examples or GitHub profile were submitted, making direct code quality assessment impossible. Based on resume descriptions alone, the candidate demonstrates senior-level system design thinking with production-grade AI pipeline architecture. A technical assessment or take-home challenge would be essential before making a final hiring decision.

+Resume descriptions suggest strong system design thinking — multi-service architecture, hybrid search with RRF, cross-encoder reranking, and SSE streaming indicate solid engineering judgment
+Experience with async patterns (Celery, async OpenSearch) and streaming (SSE) suggests awareness of performance and production-grade concerns

-No code sample, GitHub profile, or open-source work was provided — direct code quality assessment is impossible
-Cannot evaluate coding style, test coverage, documentation practices, or actual implementation quality without artifacts

Experience Overview

7y total · 4y relevant

The candidate brings 7 years of backend engineering experience with a strong 4-year trajectory into AI-integrated systems, including production LangGraph-based agentic pipelines, RAG architectures, and cloud deployments on AWS. Their current role at Cognitosoft is directly relevant and demonstrates real-world agentic AI engineering at production scale. Key gaps include missing observability tools (LangSmith, LangFuse), MCP servers, and limited evidence of ML fundamentals depth, but their overall technical profile is well-aligned with the core role requirements.

Matching Skills

PythonLangGraphOpenAI APIsVector Databases (OpenSearch)Retrieval-Augmented Generation (RAG)DockerKubernetesAWSPostgreSQLFastAPIRedisTerraformVertex AI (Gemini)Agent orchestrationNumPy (implied via ML/data work)Celery/async task processingMultimodal/LLM integrations

Skills to Verify

LangSmithLangFuseCrewAILlamaIndexSciPyMCP ServersAnthropic APIs (Claude)GitHub Actions or CI/CD pipelines (not mentioned)GCP (Vertex AI used but not GCP infra broadly)Explicit prompt engineering documentationAI observability/evaluation frameworks

Candidate information is anonymized. Personal details are hidden for fair evaluation.