CV

Download my CV or browse my experience, projects, and technical skills below.

Contact Information

Name Jacob Jerry Arackal
Professional Title AI Engineer
Email jacobjerryarackal123@gmail.com
Phone 9946994590
Location Kerala, India
Website https://jacobjerryarackal.github.io

Professional Summary

AI Engineer focused on building production-ready GenAI systems and full-stack applications that solve real business problems. Specialized in MLOps, Agentic AI, and Full-Stack development.

Experience

  • 2024 - 2025

    AI-Native Full-Stack Engineer
    Techfriar Technologies
    Led AI integration and full-stack development for enterprise clients.
    • Architected and deployed RAG pipelines and LLM agents using Gemini, OpenAI, and Llama 3.3.
    • Developed full-stack MERN applications with GraphQL and PostgreSQL for optimized data retrieval.
    • Containerized applications using Docker, ensuring 100% environment consistency between dev and prod.
    • Implemented SEO optimizations resulting in significant improvements in web app visibility.
  • 2022 - 2022

    AI/ML Engineer (Internship)
    Microsoft
    Focused on predictive modeling and cloud infrastructure.
    • Developed a deep learning model using PyTorch/TensorFlow for automobile price prediction, achieving 95% accuracy.
    • Utilized Azure ML services for experiment tracking and model deployment.
    • Performed extensive hyperparameter tuning and feature engineering to optimize model performance.

Skills

GenAI & Agentic AI (): LangChain, LangGraph, RAG, AI Agents, Multi-Agent Systems, Prompt Engineering, MCP, LLMOps, Vector Databases, Ollama
Machine Learning (): Scikit-learn, XGBoost, PCA, Random Forest, Logistic Regression, Feature Engineering, Model Evaluation
MLOps & Deployment (): MLflow, Docker, GitHub Actions, CI/CD, FastAPI, Model Serving, Experiment Tracking, Production ML Pipelines
Backend & APIs (): Python, FastAPI, Node.js, Express.js, REST APIs, GraphQL, Apollo Server
Frontend (): Next.js, React.js, TypeScript, JavaScript, HTML, CSS
Databases (): MongoDB, PostgreSQL, Pinecone, ChromaDB, SQLite, Pinecone, Redis, MySQL

Languages

English : Professional

Interests

Writing & Open Work: Publishing AI engineering content on Dev.to and LinkedIn, covering RAG, Agentic AI, LLM systems, evaluations, harness engineering, context engineering, orchestration, and MLOps.

Certificates

  • Oracle Cloud Infrastructure 2025 Generative AI Professional

Projects

  • Harness Engineering

    Autonomous engineering harness and control plane for AI-native software development with workflow routing, verification gates, architecture enforcement, runtime telemetry, and feedback-driven learning.

    • Multi-domain harness routing and progressive context assembly
    • Closed-loop telemetry and knowledge graph integration
  • MemoryOps AI

    Governed memory and context infrastructure for AI agents with policy-controlled writes, hybrid retrieval, context admission, provenance, and lifecycle management.

    • PostgreSQL RLS-enforced multi-tenant isolation
    • Deterministic governance policies and immutable audit trail
  • EvalForge

    Benchmark-driven evaluation platform for agentic systems with golden datasets, trajectory evaluation, deterministic metrics, LLM judges, experiment tracking, and regression analysis.

    • Full-trajectory observability across tool calls and latency/token budgets
    • Comparative regression delta tracking across experiment baselines
  • Multi-Agent Workflow Orchestrator

    Durable multi-agent workflow runtime with DAG scheduling, task leases, crash recovery, idempotency, evaluation gates, human approval, and audit trails.

    • Kahn’s algorithm DAG scheduling with failure recovery
    • Cryptographic SHA-256 artifact integrity and human-in-the-loop gates
  • Ad Click Prediction MLOps

    Enterprise-grade ad-click prediction pipeline with ZenML orchestration, MLflow tracking, Evidently drift detection, XGBoost training, and FastAPI serving.

    • Streamlit frontend and production API deployment
    • Sub-50ms real-time prediction target
  • Credit Card Fraud Detection

    End-to-end fraud detection MLOps pipeline with ZenML, MLflow, scikit-learn, batch inference, and an interactive Streamlit interface.

    • Precision-recall focused evaluation
    • Production-oriented deployment structure