Help & Supportcareers@gravityer.com
AI Engineer with less than a year in Generative AI & RAG systems
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Data Scientist & ML Engineer specializing in Generative AI and Agentic RAG systems, with a track record of taking multi-agent pipelines from architecture to production. Deep expertise in LangGraph orchestration, CRAG, and ReAct patterns with LLM integration across the modern GenAI stack (Gemini, Qwen, Qdrant, Pinecone, Ollama, LangChain). Proficient in Python, FastAPI, and PostgreSQL, with experience optimizing model inference efficiency and processing large-scale datasets. Distinguished by rigorous system evaluation LLM-as-judge scoring, hallucination validation, and edge case testing — delivering production-ready systems with robust security layers and real-world validation.
Helwan University in Cairo
BSc · Computers and Artificial Intelligence
August 1, 2020 – June 30, 2024
Sync Technologies
Machine Learning Intern
July 1, 2024 – September 30, 2024
India
Codsoft
Data Science Intern
June 1, 2023 – August 31, 2023
India
Podsite — AI Podcast Generation Platform
January 1, 2023 – Present
Built a full-stack AI pipeline converting scraped web content into two-speaker podcasts via a 4-stage Celery chain with real-time SSE streaming, achieving 96% valid JSON parse rate and under 2 min generation time. Synthesised 24kHz WAV output via Kokoro TTS with 94% user-rated audio clarity across 50 test episodes, reducing manual podcast production time by 80%.
CRAG Pharma-Agent
January 1, 2023 – Present
Developed a stateful Corrective RAG (CRAG) system using LangGraph and Gemini 2.5 Flash to verify clinical trials against FDA/EMA data, improving initial retrieval precision to 88%. Engineered self-correction logic with a Binary Document Grader to trigger Tavily Search fallbacks, ensuring 95% factual consistency and reducing hallucination rates by 40% for out-of-domain queries.
View ProjectPersonalized AI Training Coach
January 1, 2023 – Present
Developed a 4-agent LangGraph system to generate adaptive periodization plans based on wearable data (HRV, sleep) and RAG-driven retrieval over 342 PubMed articles indexed in Qdrant. Engineered an automated feedback loop monitoring ACWR and HRV signals to dynamically trigger progression or deload weeks. LLM-as-judge score using Llama 3.3 70B rating 20 plans on 5 criteria and Correct trigger detection on 11 edge cases with 0 false positives.
Athletic Performance Intelligence System
January 1, 2023 – Present
Architected a 9-node Agentic RAG pipeline utilizing hybrid search (BM25+dense) across 3 Pinecone namespaces, integrating automatic Tavily web fallbacks for low-confidence retrievals (¡0.50 score). Engineered a 5-dimension Gap Detector with up to 3 reflexion retries for targeted web re-queries, driving a streaming DeepSeek generator with strict citation hallucination validation. Implemented robust security and routing layers, including a 2-tier prompt injection defense, Redis rate-limiting (5 req/min), and LLM routing (DeepSeek/Llama-3.3) achieving~400 TPS.
Cultural Fit Analysis
The candidate's portfolio showcases a strong passion for cutting-edge AI technologies, particularly in Generative AI and Agentic RAG systems. The personal projects are highly relevant to an AI Engineer role, demonstrating initiative and self-driven learning beyond academic requirements. The diversity of projects, from personalized training coaches to pharma agents and podcast generation, indicates a broad interest in applying AI across different domains. The candidate's focus on robust system evaluation and production-readiness aligns well with a culture that values quality and reliability in AI deployments. The candidate's academic background in Computers and AI further reinforces a strong foundational fit for technical roles in AI.
Soft Skills & Operational Fit
The candidate's project descriptions highlight a strong problem-solving aptitude, particularly in addressing complex challenges like hallucination reduction, prompt injection defense, and dynamic feedback loops. The focus on quantifiable results (e.g., 0 false positives, 95% factual consistency, 400 TPS) indicates a results-oriented and detail-conscious approach. The ability to architect multi-agent systems suggests strong organizational and system-thinking skills. While direct collaboration experience is not explicitly detailed, the complexity of the projects implies an ability to manage intricate technical tasks independently.