Divpreet Singh
AI Engineer with 6+ years of experience building scalable backend systems and production-grade AI/LLM pipelines. Specializing in RAG applications, entity resolution systems, AI-driven data processing, and LLM orchestration workflows for financial and enterprise environments.
Featured Work
Production AI Projects
End-to-end AI systems built for scale, reliability, and real-world impact.
Entity Resolution & AI Data Platform
High-Scale Transaction Normalization
Processing, normalizing, grouping, and deduplicating noisy financial transaction narration inputs across 300K–1.8M+ records with high precision.
View case study: Entity Resolution & AI Data Platform
Key Features
- AI-integrated entity extraction and resolution platform for financial datasets
- High-throughput pipelines for normalization, grouping, and deduplication
- Optimized workflows handling 300K–1.8M+ records with low retrieval latency
- Used embeddings, semantic similarity, and Ray for scalable matching
Architecture
Raw Data
Financial narrations
AI Extractor
Entity resolution
Ray Cluster
Parallel matching
Elasticsearch
Indexed storage
Deduplication
Normalized output
AI Financial Question Answering (RAG)
Contextual Grounding & Retrieval Evaluation
Accurate retrieval, contextual ranking, and factual grounding for financial document QA with hallucination protection and retrieval evaluation.
View case study: AI Financial Question Answering (RAG)
Key Features
- RAG-based financial QA system using structured LLM workflows
- Chunking, embeddings, and semantic search optimization
- Retrieval augmentation, contextual re-ranking, and orchestration workflows
- Retrieval evaluation, response grounding, and factual validation
Architecture
Query
User financial query
Chunker
Document parser
FAISS Index
Vector search
Re-ranker
Contextual ranker
LLM Answer
Grounded response
Agent Evaluation & AI Workflow Platform
Correctness & Hallucination Detection
Automated evaluation, hallucination detection, and prompt scoring for production LLM applications and agentic workflows.
View case study: Agent Evaluation & AI Workflow Platform
Key Features
- Evaluation workflows for LLM correctness, grounding, and hallucination detection
- Automated pipelines for prompt evaluation, response scoring, and testing
- Validation workflows for comparing prompts, models, and AI-generated outputs
- Explored agent-oriented execution and evaluation frameworks
Architecture
Prompt/Input
Test dataset
LLM Runner
Model execution
Evaluator
Grounding check
LangSmith
Trace tracking
Metrics
Accuracy score
Multi-Agent Voice Support System
Real-Time Conversational AI Pipeline
Orchestrating end-to-end voice support with speech-to-text, knowledge retrieval, LLM reasoning, and text-to-speech agents.
View case study: Multi-Agent Voice Support System
Key Features
- Multi-agent voice-based customer support platform (STT + Retrieval + Reasoning + TTS)
- Agent orchestration workflows for customer query handling & contextual response
- Integrated retrieval-based knowledge access with LLM generation
- Modular architecture supporting scalable deployment and future agent extensions
Architecture
Audio In
Voice stream
Whisper STT
Transcribe audio
Agent Router
Reason & retrieve
LLM Core
Response generation
TTS Out
Audio response
OCR & Intelligent Document Processing
Layout-Aware Extraction & Parsing
Automated layout-aware extraction, table parsing, and classification for noisy scanned financial documents, PAN, and Aadhaar cards.
View case study: OCR & Intelligent Document Processing
Key Features
- OCR workflows for PAN, Aadhaar, and financial document processing
- Improved extraction quality from noisy and scanned inputs
- Developed preprocessing and structured parsing logic
- Document classification, layout-aware extraction, table parsing, and validation
Architecture
Scanned Doc
Input image/PDF
Preprocess
OpenCV cleanup
OCR Engine
Tesseract extraction
Layout Parser
Table & field map
JSON Result
Structured output
AI Infrastructure & Governance System
Resource Allocation & Workload Optimization
Optimizing CPU/memory-intensive AI workloads and managing Kubernetes/Docker deployment infrastructure.
View case study: AI Infrastructure & Governance System
Key Features
- Worked on optimizing CPU/memory-intensive AI workloads
- Developed resource allocation and monitoring workflows
- Supported scalable execution of backend processing systems and GPU environments
Architecture
Workload
AI/ML job queue
Governor
Resource allocator
K8s Nodes
Auto-scaling pods
AWS/GCP
Cloud execution
Monitor
Metrics & logs
Career
Work Experience
Building AI-powered backend systems at scale across fintech, e-commerce, and enterprise SaaS.
Software Engineer (AI/Backend)
- Designed & implemented AI-driven entity extraction and resolution pipelines for financial transaction data, scaling to 300K–1.8M+ records
- Implemented parallel processing using Ray and async execution, reducing batch pipeline execution time and boosting system performance by 30–35%
- Integrated LLM-based components, RAG-based retrieval systems, and LLM orchestration workflows into backend services for financial query answering
- Developed & optimized high-volume Elasticsearch pipelines and financial narration normalization pipelines
- Deployed GPU-enabled Docker environments on GCP and FastAPI backend microservices for ML/AI workloads
Software Developer
- Built robust backend services and RESTful APIs for data analytics workflows in an Agile environment
- Designed high-throughput ETL data processing pipelines for large-scale datasets
- Improved API performance and query response times through database indexing and async execution
- Delivered production features end-to-end with high code quality and test coverage
Software Developer
- Developed scalable backend APIs using Python and SQL for core business applications
- Implemented secure authentication, authorization, and input validation workflows
- Assisted in CI/CD pipeline automation and rapid production issue resolution
Expertise
Technical Skills
Deep expertise across AI/ML, backend engineering, and distributed systems.
AI / ML & RAG Design
AI Engineering & Security
Backend & Systems
Search, Data & Processing
DevOps, MLOps & Cloud
Insights
Technical Blog
Deep dives into AI engineering, system design, and production lessons learned.
Get in Touch
Let's Connect
Interested in AI engineering, RAG pipelines, backend systems, or potential opportunities? Reach out directly!
divpreet1012@gmail.com
GitHub
github.com/Divpreetai
linkedin.com/in/divpreet-singh12
Resume
Download PDF
Have an opportunity in mind?
Tell me about your team, project, or role and what you're looking to build. I'd love to hear from you.
Email meOpens a draft in your email app. You can also copy the address below into your preferred email service.
divpreet1012@gmail.com
Based in Mohali, India