Kaushalendra Pratap Singh

AI Engineer

About Me

AI/ML Engineer building production-grade machine learning and GenAI systems. Strong background in NLP, generative AI (LLMs, embeddings, semantic search), and Python-based ML development. Hands-on experience with model deployment, MLOps, evaluation frameworks, and scalable cloud architectures. Proven ability to translate business requirements into reliable, enterprise-grade AI solutions with focus on responsible AI practices.

Building production-grade GenAI systems (RAG, agents, document intelligence)

Kaushalendra Pratap Singh

The Specs

Location: Gandhinagar, Gujarat, India
Email: kaushalendrapratapsingh92@gmail.com
Education: B.Tech. in Computer Science & EngineeringSRM University8.56 / 10.0
Certification: AWS Certified Developer – Associate

Tech Stack

Machine Learning & AI

Scikit-learnPyTorchTensorFlowProphetMLflowModel EvaluationFeature EngineeringDrift Monitoring

NLP & Transformers

Hugging Face TransformersBERTTokenizationSemantic SearchSimilarity SearchText Embeddings

GenAI & LLMs

GPT-4oClaudeGroqAzure OpenAIRAG PipelinesEmbeddings (FastEmbed)Prompt EngineeringLangChainGuardrails

Vector Databases

FAISSChromaDBOpenSearchRetrieval Tuning (top-k, MMR)Semantic Chunking

Backend & APIs

PythonFastAPIFlaskNext.jsREST APIsAsync Processingn8n Automation

Data & Storage

MySQLPostgreSQLNoSQL (Redis)AirtableData PreprocessingETL Pipelines

Cloud & MLOps

AWS (EC2, Lambda, S3)GCPAzureDockerCI/CD (GitHub Actions)Model MonitoringExperiment Tracking

Responsible AI

Privacy ControlsPII-safe LoggingEvaluation MetricsOutput ValidationCompliance Awareness

How I Design AI Systems

Read

From PDFs to Actions

Building a Document Intelligence Automation Workflow with GenAI & n8n. End-to-end RAG platform with structured extraction and DocOps automation.

Blog Soon

Evaluating LLM outputs

Hallucination detection through structured outputs, confidence scoring, and automated fact-checking against source documents.

Blog Soon

Cost vs latency trade-offs

Batching requests, intelligent caching, model routing (small models for simple tasks), and async processing to balance performance with budget.

Blog Soon

Why agents need guardrails

Tool validation, retry logic with exponential backoff, output sanitization, and human-in-the-loop for high-stakes decisions.

Beyond the Code

Running & Fitness

Training for an official marathon in 2026. Runner-up at Yoga Asana Championship and Fit4Life 10K.

Sports

Basketball • Badminton • Pool (Billiards)

Reading

Crime Thrillers • Biographies • Sci-Fi • Horror

Entertainment

F.R.I.E.N.D.S series • Harry Potter movies • Music (all day!)