01about
From Full-Stack to AI Engineering
I started as a full-stack engineer building with React, Node.js and Go — shipping APIs, pipelines and dashboards that had to be fast, reliable and observable. That production discipline is the foundation of everything I do now.
Over the past three years I pivoted into data science and AI engineering: training and fine-tuning models, building RAG pipelines, and operationalizing ML in the same rigorous way I once shipped web services.
Currently focused on Large Language ModelsMLOps & Production MLPredictive Modeling.
career.log
[2018] FULL-STACK
React / Node / Go — shipped 20+ products
[2022] MLOPS
Production ML, drift monitoring, pipelines
[2024] AI ENGINEERING
LLMs, RAG, fine-tuning at scale
tail -f next_chapter
02skills
Technical Arsenal
ali@dev:~$ toolchain --list
LANGUAGES
- Python
- R
- SQL
- TypeScript
- Go
- Bash
ML / DL
- PyTorch
- TensorFlow
- scikit-learn
- LangChain
- Hugging Face
DATA
- Pandas
- Polars
- Apache Spark
- dbt
- Airflow
DEPLOYMENT
- Docker
- Kubernetes
- FastAPI
- Terraform
- AWS
03certifications
Credentials & Track Record
verified credentials
01
Google Professional Machine Learning Engineer
Google Cloud · 2025
ID: mock-id-GCP-MLE-2025
02
AWS Certified Machine Learning — Specialty
Amazon Web Services · 2024
ID: mock-id-AWS-ML-2024
03
DeepLearning.AI LLM Certification
DeepLearning.AI · 2024
ID: mock-id-DLAI-LLM-2024
04
TensorFlow Developer Certificate
Google · 2023
ID: mock-id-TFD-2023
measurable impact
impact.log
- Built an AutoML pipeline that reduced model drift by 40% across 12 production models.
- Deployed a RAG pipeline serving 1M queries per day at < 350ms p95 latency.
- Cut inference costs by 52% by quantizing and caching LLM serving stacks.
- Led a 4-person team shipping a churn-prediction platform with 91% precision.
04projects
Selected Work
PROJECT-01
1M+ queries/day
NeuralRAG
Production RAG pipeline with hybrid retrieval (dense + BM25), reranking, and guardrails. Handles 1M+ queries/day with streaming responses.
- Python
- LangChain
- FastAPI
- Qdrant
- Docker
PROJECT-02
-40% model drift
DriftGuard AutoML
Automated feature engineering, model selection and drift monitoring. Cut model drift by 40% across 12 production deployments.
- Python
- PyTorch
- Airflow
- Kubernetes
- Evidently
PROJECT-03
-52% inference cost
LLM Observability Stack
Open-source toolkit tracing prompt, token and cost metrics across LLM calls. Plugs into any OpenAI-compatible endpoint.
- Go
- TypeScript
- OpenTelemetry
- ClickHouse
PROJECT-04
91% churn precision
ForecastEngine
Time-series forecasting service with Prophet + gradient boosting ensembles, feature stores and API-first design.
- Python
- Polars
- FastAPI
- PostgreSQL
- Docker
05kaggle & open source
Competitions & Contributions
Competitions Expert
Global rank #1421
Titanic — Machine Learning from Disaster
Silver
Top 5%
House Prices — Advanced Regression
Silver
Top 8%
Spaceship Titanic
Bronze
Top 12%
CommonLit Readability Prize
Bronze
Top 15%
merged pull requests
pandas-dev/pandasGH-48213
Merged: faster groupby aggregation for categorical columns (2.3x speedup).
langchain-ai/langchainGH-17104
Merged: multi-query retriever improvements with configurable search modes.
apache/sparkGH-39412
Merged: optimized window function partition cleanup (18% memory reduction).
06contact
Let's Build Something Intelligent
I'm currently open to senior AI/ML engineering roles — building LLM infrastructure, MLOps platforms, or production data science. If your team is solving interesting problems, let's talk.