About
Experience
Projects
Skills
Contact

ABOUT

Building production-grade AI systems powered by Graph-RAG, Agentic AI, and scalable backend engineering.

I'm Anupam Kumar, an AI Engineer and Computer Science graduate from IIITM Gwalior. My focus is building intelligent systems that combine large language models, retrieval pipelines, autonomous agents, and modern software engineering practices.

Over the past few years I've worked across Graph-RAG, multi-agent architectures, computer vision, backend systems, and MLOps infrastructure. I enjoy taking ideas from research papers and turning them into reliable products that can run in production environments.

Core Areas

Agentic AIGraph-RAGMulti-Agent SystemsLangGraphKnowledge GraphsFastAPIComputer VisionMLOpsNeo4jVector SearchAWSDocker

EXPERIENCE

AI Engineer

Yahweh Innovations — AI Solutions & Consulting

Remote

Apr 2025 – Dec 2025
  • ✦Built and deployed a production-grade chatbot and knowledge retrieval platform serving 10,000+ monthly requests across enterprise clients.
  • ✦Reduced average query resolution time from 45 seconds to 29 seconds through retrieval optimization and backend performance improvements.
  • ✦Automated customer support, onboarding, and audit workflows processing 500+ tasks daily while reducing manual effort by 55%.
  • ✦Delivered an OCR-powered real-estate audit platform achieving 95%+ extraction accuracy and reducing processing time from days to minutes.

Independent AI Consultant

Freelance AI & Automation Projects

Remote

May 2024 – Jul 2024
  • ✦Developed custom AI automation solutions and backend systems for small businesses and independent clients.
  • ✦Built data processing pipelines, REST APIs, and workflow automation tools tailored to client requirements.
  • ✦Managed end-to-end project delivery including solution design, implementation, deployment, and client support.

Featured Projects

Swipe to explore my latest work

AetherCV

Production AI Research Platform for Computer Vision

End-to-end AI platform that combines semantic routing, hybrid retrieval, graph-enhanced reasoning, observability, evaluation pipelines, and production infrastructure to deliver evidence-grounded answers over computer vision research literature.

Key Achievements
Hybrid retrieval system combining FAISS, BM25, RRF Fusion, and Cross-Encoder Reranking
0.94 RAGAS Context Recall, 0% hallucination rate and 100% grounded-answer benchmark performance
Seven-layer intelligent caching architecture using Redis and semantic cache matching
PythonFastAPIFAISSRedisPostgreSQLPrometheusGrafanaMLflowDockerNginxGroqRAGAS
Code
AetherCV
94%
RAGAS Recall

LeadBoost AI

Production AI Lead Intelligence SaaS

Full-stack AI platform that automates lead discovery, enrichment, qualification, and personalized outreach through distributed processing, LLM-powered workflows, and scalable SaaS infrastructure.

Key Achievements
Built multi-tenant SaaS architecture with authentication, subscriptions, and RBAC
Engineered Celery + Redis distributed task queues for large-scale lead processing
Automated lead enrichment and AI-powered prospect qualification workflows
FastAPICeleryRedisPostgreSQLDockerReactLangChainPlaywright
Live DemoCode
LeadBoost AI
80%
Manual Effort Reduced

Ayurgenix AI

Agentic AI Platform for Ayurvedic Knowledge Systems

Production-grade AI platform that transforms classical Ayurvedic literature into an intelligent knowledge system using multi-agent reasoning, semantic retrieval, and grounded LLM generation.

Key Achievements
Built an LLM-powered retrieval platform over 10,000+ medical knowledge documents, validated across 500+ evaluation queries using RAGAS-based evaluation frameworks
Implemented agentic retrieval workflows with query expansion, intent classification, reranking, and confidence-based response generation
Engineered Pinecone-powered semantic search achieving sub-second retrieval latency with improved context relevance and citation quality
FastAPIPineconePostgreSQLDockerPyTorchSentence TransformersLangSmithJWT
Live DemoCode
Ayurgenix AI
<500
ms Response Time

TalentForge AI

AI-Powered Job Application Orchestration Platform

End-to-end intelligent automation platform that combines LLM-powered career intelligence, browser automation, analytics, and workflow orchestration to streamline large-scale job application management.

Key Achievements
Designed a multi-layer architecture with Controller, Service, Storage, Platform, and Analytics components
Built LLM-based resume compatibility scoring and intelligent application prioritization workflows
Automated application execution using Playwright with anti-detection mechanisms and safety controls
PythonGroqPlaywrightStreamlitSQLitePandasAsyncIODocker
Live DemoCode
TalentForge AI
90%
Manual Effort Reduced

SentinelAI IDS

Real-Time AI-Powered Network Threat Detection Platform

Production-grade intrusion detection system that combines machine learning, anomaly detection, explainable AI, and real-time traffic analysis to identify cyber threats across enterprise networks.

Key Achievements
Built a hybrid detection pipeline combining Random Forest classification and Autoencoder-based anomaly detection across 39 network flow features
Implemented real-time packet capture, flow aggregation, and threat analysis using Scapy and WebSocket-based streaming architecture
Detected DDoS, Botnet, Brute Force, Infiltration, Port Scan, and Web Attack patterns with explainable AI insights powered by LIME
PythonTensorFlowScikit-LearnFlaskReactNode.jsMongoDBScapyWebSocketsLIME
Code
SentinelAI IDS
95%
Detection Accuracy

View More

Check out more projects on GitHub

Visit GitHub

Skills & Expertise

Designing intelligent systems that combine LLMs, retrieval architectures, backend platforms, observability, and cloud infrastructure for production environments.

LLM Engineering

Building production LLM applications with retrieval, agent workflows, semantic search, model serving, and orchestration frameworks.

LangChainLangGraphCrewAIOpenAITransformersvLLMOllama

Retrieval Systems

Designing knowledge systems powered by hybrid retrieval, reranking, vector search, graph reasoning, and citation-grounded responses.

Neo4jFAISSPineconeBM25Hybrid SearchRRF FusionSemantic RouterRAGAS

AI Platform Engineering

Developing scalable backend systems, APIs, microservices, and workflow automation platforms for production AI applications.

FastAPIPostgreSQLRedisCelerySQLAlchemyREST APIsWebSocketsMicroservices

Computer Vision

Building OCR pipelines, document intelligence systems, image processing workflows, and deep learning applications.

PyTorchOpenCVOCRComputer VisionYOLOCNNsImage ProcessingDeep Learning

Cloud Infrastructure

Deploying containerized applications with CI/CD, cloud-native architectures, automation workflows, and scalable infrastructure.

DockerAWS EC2GitHub ActionsLinuxCI/CDNginxCloud Deployment

MLOps & Observability

Ensuring reliability through monitoring, evaluation, experimentation, telemetry, and continuous improvement pipelines.

MLflowPrometheusGrafanaRAGASModel EvaluationMonitoringTelemetry
// GET IN TOUCH

Let's Connect

Open to AI engineering roles, Graph-RAG consulting, and technical collaborations.

Email

anupraj1620@gmail.com

GitHub

github.com/anupamkr1708

LinkedIn

linkedin.com/in/anup-kumar

Resume

Download CV