AI Research Engineer (Early Career)

1.5y relevant experience

Qualified

Executive Summary

The candidate is a strong entry-level candidate who significantly outpaces the typical new-grad applicant for this role. Their 1.5 years at Oracle shipping production RAG systems, agent memory pipelines, and conversational AI features — combined with a research internship implementing graph algorithms for GraphRAG — directly maps to what this role requires day-to-day. The primary technical gap is LangGraph, and the primary process gap is the absence of any verifiable code (no GitHub, no submission). These gaps are material enough that a technical screen is essential before advancing, but if they performs well on a hands-on assessment, they represents excellent value within the stated salary range and strong upside for a high-mentorship, fast-moving team.

Top Strengths

✓Production RAG and AI agent memory systems built and shipped at Oracle — directly matches core job responsibilities
✓Systems-level thinking demonstrated: embedding quality optimization, chunking architecture, retrieval pipeline design at an enterprise scale
✓Research credibility: implemented Leiden algorithm from a research paper at Oracle Labs PGX, showing ability to translate academic ML work into code
✓Strong polyglot engineering background (Python, Java, C, PL/SQL) with exposure to multiple vector DB platforms (ChromaDB, Pinecone, Neo4j)
✓Trilingual (Arabic native, English fluent, French B2) and demonstrably capable of async written communication (blog posts, masterclasses) — well-suited for remote-first culture

Key Concerns

!No submitted code sample or public GitHub makes direct engineering quality verification impossible — a critical gap before making a hiring decision
!Missing LangGraph experience specifically, which is listed as a required skill for agentic workflow development — the central technical domain of this role

Culture Fit

74%

Growth Potential

High

Salary Estimate

$45,000 - $58,000 (entry-level band, Morocco-based, likely flexible given Oracle experience adding leverage)

Assessment Reasoning

FIT decision is driven by the candidate's unusually direct match to the core job responsibilities: they have shipped production RAG pipelines, designed AI agent memory systems, and contributed to retrieval architecture at Oracle — a Fortune 100 engineering environment — while still early in their career. They meets approximately 6 of 9 listed required skills (Python, LLM, RAG, AI Agents, vector databases, prompt engineering), with Java as a bonus systems language, and is missing only Rust and LangGraph. Their academic ML project work (AutoNoSQL cost optimization, Spark ML pipeline) and research experience (Leiden algorithm implementation) demonstrate the intellectual depth and engineering rigor the role values. The FIT designation carries a confidence penalty (72%) due to two unresolved gaps: (1) no code sample or public GitHub, leaving engineering quality unverifiable, and (2) missing LangGraph which is central to the agentic workflow responsibilities. A technical screen or short take-home exercise is strongly recommended as a prerequisite to advancing — if they performs well, they should be considered a priority hire within the entry-level band.

Interview Focus Areas

Deep technical dive on Oracle RAG architecture: chunking strategy decisions, embedding model choices, retrieval evaluation — probe for genuine ownership vs. surface-level involvementLangGraph and agentic orchestration: assess current knowledge gap, learning speed, and how quickly they could become productive with LangGraph given their agent memory backgroundProduction engineering discipline: versioning, observability, cost/latency tradeoffs — key role requirements they haven't explicitly addressedAsync remote work self-direction: scenarios testing independence, communication cadence, and comfort with ambiguity in a small team

Code Review

FairJunior Level

No code example or GitHub profile was provided, preventing direct assessment of code quality, style, or engineering discipline. The resume narrative describes production-level engineering work at Oracle with reasonable specificity, suggesting competence above a typical new grad, but this cannot be verified without seeing actual code. A technical screen or take-home exercise is strongly recommended before advancing this candidate.

PythonJavaPL/SQLCScikit-learnDockerKafkaSpark MLMongoDB

+Resume descriptions of technical work imply structured, production-aware engineering — chunking strategies, pipeline design, and API development at Oracle suggest code quality above average for early career
+Academic project (AutoNoSQL) shows cost-aware engineering mindset: classifier gating LLM calls to reduce expensive inference by ~50%

-No code sample was submitted and no public GitHub profile was provided, making direct code quality assessment impossible — this is a meaningful gap for an engineering role

Experience Overview

1.5y total · 1.5y relevant

The candidate presents an unusually strong profile for an entry-level candidate, with 1.5 years of real production experience at Oracle shipping RAG pipelines and AI agent memory systems into a live enterprise product. Their technical choices — chunking strategy optimization, embedding quality improvements, semantic search design — reflect the kind of systems thinking the role explicitly seeks. The main gaps are Rust and LangGraph, which are listed requirements, but their Java, Python, and C background plus demonstrated ability to learn from papers suggests these are bridgeable.

Matching Skills

PythonLLMRAGvector databasesprompt engineeringJavaAI Agentsembeddings

Skills to Verify

RustLangGraphexplicit fine-tuning experience

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