Do I need a large dataset to use ML?

Depends on the approach. Custom models typically need hundreds to thousands of labeled examples. LLM-based systems and RAG pipelines can work with your existing documents with minimal prep.

How do you handle model accuracy and testing?

At Creative Codes, every model we ship comes with documented precision and recall scores on held-out test data. We define what 'good enough' means with you upfront and don't deploy until the model hits that bar.

Can you integrate with our existing systems?

Yes. We deploy behind REST or GraphQL APIs, as serverless functions, or as containers that plug into your existing infrastructure. We match your stack.

What about ongoing model maintenance?

Models drift as data changes. We set up monitoring that catches accuracy drops and triggers retraining automatically. You don't have to watch it.

How is this different from using ChatGPT directly?

At Creative Codes, we build constrained systems with RAG, fine-tuning, and structured outputs that are reliable enough to automate against. ChatGPT hallucinates, can't access your private data, and isn't tuned for your specific task. Our systems are.

What does an AI or ML project typically cost?

Most AI and ML projects range from $3,000 to $15,000 depending on model complexity, data volume, and integration requirements. RAG pipelines for document Q&A typically fall in the $4,000 to $8,000 range. All projects are fixed-price with code in your GitHub.

Services · AI & Machine Learning

AI & Machine LearningModels that work in production,not just notebooks.

Custom ML models, LLM integrations, and RAG pipelines built for production data at production volume. No science projects.

Book a discovery call View case studies

97%+ production model accuracy10 ML systems shippedTop Rated Plus on Upwork

Need raw data first? Web scraping services →

How engagements work →

Fixed-price builds. Code in your GitHub from day one.

→ github.com/hassan1731996

Featured work

→ Enterprise RAG Knowledge System → MiningCave AI Customer Agent

analyze.py · bashrunning

# NLP analysis: 10 000 product reviews

$ python analyze.py --input reviews.jsonl --model sentiment-v3

✓ model loaded distilbert-base (67M params)

✓ vectorstore 148 402 embeddings indexed

→ [0000/10000] processing batch…

✓ [2500/10000] acc 97.2% | 840 docs/s

✓ [5000/10000] acc 97.1% | 855 docs/s

✓ [7500/10000] acc 97.4% | 861 docs/s

✓ [10000/10000] done | 1.3s total inference

→ 1 284 positive · 8 341 neutral · 375 negative

→ 47 anomalies flagged → automation layer

94%

query accuracy on production RAG systems

82%

support queries handled without a human

<1%

hallucination rate on grounded LLM responses

What we build

Every layer of the ML stack.

Custom ML Models

Classification, regression, clustering. Trained on your data, tuned to your KPIs, served behind a FastAPI endpoint.

LLM Integration & Fine-Tuning

Connect GPT, Claude, or open-source models to your workflows. Fine-tune for domain-specific accuracy when general models fall short.

RAG Pipelines

Retrieval-augmented generation that grounds LLM responses in your actual data. Vector search, chunking strategy, and hallucination reduction.

NLP & Text Analysis

Sentiment analysis, entity extraction, document classification, and summarization. One document or 50k, same interface.

Anomaly Detection

Statistical and ML-based detection of outliers in pricing, transactions, user behavior, or equipment telemetry. Real-time or batch.

Predictive Analytics

Forecast demand, churn, pricing trends, or inventory needs. Models retrained automatically as new data flows in.

Need a model or RAG pipeline built?

Tell us your data and the decision you need to make. We scope it after a free call.

Book a discovery call

Stack

PythonClaudeOpenAI APILangChainLangGraphLlamaIndexOpenClawPyTorchHugging FaceQdrantOllamaFastAPI

How the code looks.

pipeline.py

async def process_reviews(reviews: list[str]) -> AnalysisResult:
    embeddings = model.encode(reviews, batch_size=64)
    clusters = HDBSCAN(min_cluster_size=10).fit(embeddings)

    sentiments = await asyncio.gather(*[
        analyze_sentiment(review) for review in reviews
    ])

    topics = extract_topics(embeddings, clusters.labels_)
    anomalies = detect_outliers(sentiments, threshold=2.5)

    return AnalysisResult(
        topics=topics,
        sentiment_distribution=Counter(sentiments),
        anomalies=anomalies,
        cluster_count=clusters.labels_.max() + 1
    )

model.encode(reviews, batch_size=64)

Sentence-transformer embeddings. Batching avoids OOM on large review sets. 64 is tuned for a 16GB GPU.

HDBSCAN(min_cluster_size=10)

Density-based clustering that discovers the number of clusters automatically. No need to pre-specify K.

detect_outliers(sentiments, threshold=2.5)

Z-score based anomaly detection on the sentiment distribution. 2.5σ catches ~1.2% of reviews for manual review.

After we process your data.

Insights without action are just reports. Our automation layer turns ML outputs into triggered workflows: alerts, updates, API calls, and business actions that happen without human intervention.

Process

Act

See Automation capabilities →

17 reviews5.0 avg100% Job Success on Upwork

From clients

Top Rated Plus · 100% Job Success · $50K+ earned

Exceptionally skilled back-end developer. Deep technical expertise in refactoring complex systems and building scalable multi-tenant architectures. Responsive, proactive, and consistently delivered above expectations.

Turki Alelyani

Founder, Feelix AI LLC, United States

5.0Upwork

Professional, responsive, and clearly committed to high quality work. Asked smart questions up front, provided progress updates without being asked, and delivered exactly what I needed on time.

Steven Cohen

GreenMark Consulting Group, United States

5.0Upwork

Hassan is responsive, detail-oriented, and thorough. He introduced AI combined with telecom into our projects and the results have been strong.

Sean Kannegiesser

IT / MSP Manager, Canada

See all 17 reviews