Available for AI Engineer, Data Scientist & Data Analyst roles ·

Building
agentic systems
that ship.

I'm Mohammed Ajzel — an AI/ML Engineer and Data Scientist turning complex data into production systems. I build Agentic AI, RAG pipelines, fine-tuned LLMs, and analytics dashboards that drive measurable outcomes.

LocationBangalore, IN
FocusAgentic · RAG · LLMs · Data Science
Status● Open to AI & Data roles
LangGraphMCPRAGFAISSLoRA / QLoRAUnslothPyTorchPandasNumPySQLTableauPower BIFastAPIDockerOCI GenAI CertifiedAnthropic ClaudeGroqOllamaMatplotlib
LangGraphMCPRAGFAISSLoRA / QLoRAUnslothPyTorchPandasNumPySQLTableauPower BIFastAPIDockerOCI GenAI CertifiedAnthropic ClaudeGroqOllamaMatplotlib
01 / Outcome
70%
Efficiency gains
n8n automation @ Wayfair
02 / Outcome
92%
Model accuracy
Random Forest, R² @ IBM
03 / Outcome
0.82
ROC-AUC
Churn model, Lloyds
04 / Outcome
85–90%
RAG accuracy
FAISS + Gemini pipeline
§ 01 · What I do

From data to intelligence — across AI, analytics, and the full stack.

01Discipline

Agentic AI Systems

Multi-agent architectures with LangGraph, LangChain, and the Model Context Protocol. ReAct reasoning across tool servers, deployed as FastAPI services.

LangGraphMCPLangChainReAct
02Discipline

Retrieval-Augmented Generation

Production RAG pipelines with FAISS vector search, dual embeddings, and sub-1.5s query latency across multi-document corpora.

FAISSGeminiMiniLMStreamlit
03Discipline

LLM Fine-Tuning

LoRA / QLoRA fine-tuning with Unsloth on Phi-3-mini. 4-bit quantized training, GGUF export, offline inference via Ollama.

LoRAQLoRAUnslothGGUF
04Discipline

ML & Data Science

Statistical modeling, EDA on 10k+ records, A/B testing, and classification pipelines with measurable business outcomes. SQL, Pandas, NumPy, Scikit-learn.

PyTorchSklearnPandasNumPySQL
05Discipline

Data Analytics & Visualization

End-to-end analytics workflows: data cleaning, feature engineering, dashboarding, and storytelling with Tableau and Power BI. Translate raw data into board-ready insights.

TableauPower BISQLMatplotlibSeaborn
§ 02 · Selected work

AI systems & data pipelines.
Real users. Real numbers.

MCP-01

MCP LangChain Agentic System

Python · MCP · LangGraph · Groq · Docker
Multi-server MCP agentic system routing queries to a LangGraph ReAct agent (Groq/Qwen-32B). Dockerized full stack cut setup time 80% and logs every interaction for downstream analysis.
2+ tool servers · 80% faster deploys
LLM-02

LLM Fine-Tuning Pipeline

Unsloth · LoRA/QLoRA · Phi-3-mini · Ollama
Fine-tuned Phi-3-mini (4-bit) on HTML→JSON extraction. LoRA r=64, α=128, SFTTrainer with 8-bit AdamW. Tracked experiments with loss curves and validation metrics for reproducible tuning.
60% memory reduction · Zero cloud
RAG-03

AI Document Searchbot

LangChain · FAISS · Gemini · FastAPI
RAG pipeline with dual embeddings (Gemini + MiniLM offline) achieving 85–90% conversational QA accuracy across 10+ PDFs. Evaluated with precision@k and mean response latency.
85–90% accuracy · <1.5s latency
ATS-04

Cloud-Native ATS Analyzer

Streamlit · Gemini · Docker · Render
Gemini-powered semantic ATS matching resumes to job descriptions. Built analytics dashboards for match scores, skill gaps, and screening funnel metrics. Sub-2s inference.
60% screening reduction · <2s
DATA-05

Data Science & Analytics Portfolio

Python · Pandas · SQL · Tableau · Power BI
End-to-end data projects covering EDA, feature engineering, statistical testing, predictive modeling, and executive dashboards. Clean datasets, reproducible notebooks, and actionable insights.
EDA → model → dashboard
CNN-06

Deepfake Image Detection

PyTorch · CNN · Flask · Docker
PyTorch CNN with data-augmentation pipeline hitting 92% validation accuracy at <50ms inference. Real-time Flask app serving 100+ users.
92% accuracy · <50ms inference
§ 03 · Trajectory

Four internships. Every one shipped something real.

2026 · 01 → 0301 / 04

AI Automation Extern

Wayfair (via Extern)

n8n pipeline with Claude + Gemini across 48+ SKUs — 70% less manual work. Produced a 19-page intelligence report with structured data insights, cutting turnaround to <2h.

2026 · 02 → 0402 / 04

AI Engineer Intern

Infosys Springboard 6.0

Enterprise AI email categorization + prioritization with transformer models. Analyzed ticket metadata and built prompt-engineered classification with measurable accuracy gains. Certified by Infosys SVP.

2025 · 12 → 26 · 0103 / 04

Machine Learning / Data Analyst Intern

Cognifyz Technologies

EDA + statistical benchmarking on 10k+ records; Random Forest at 90%+ accuracy. Built a Flask recommendation API with Top-20 personalized results and documented feature importance for stakeholders.

2025 · 05 → 0704 / 04

AI, ML & Data Science Intern

IBM

Flask energy prediction API (Random Forest, 92% R²) at 100+ predictions/sec. Created a Chart.js real-time dashboard for monitoring predictions and residuals, bridging modeling with analytics.

Certifications
Oracle OCI Generative AI Professional
Anthropic — Claude Code 101 & In Action
IBM Data Analyst
Google Cloud — GenAI for Data Analytics
Infosys — LLMs for Developers