Available for AI & Backend Roles

ANHAR
ESWARAMANGALAM

AI & Data Science Engineer

Bridging mathematical machine learning and low-latency production reality. Architecting bilingual RAG pipelines in English & Malayalam, containerized microservices, and autonomous LLM agents designed to scale under load.

Explore Production Work
Connect:GitHub•LinkedIn•Email
Anhar Eswaramangalam - AI & Data Science Engineer
FastAPI // Async Core
pgvector // HNSW Index
Agno // Bilingual RAG
Portfolio Build · 2026.11•Next.js 15 · Three.js · Framer Motion
Scroll to explore
Executive Profile

Ibridgethegapbetweenexperimentalmachinelearningandproduction-ready,low-latencysoftware.GroundedinarigorousBachelorofComputerApplicationsfoundation,IarchitectbilingualRAGpipelines,containerizedFastAPImicroservices,andautonomousLLMagentsdesignedtoscale.

I am an AI & Data Science Engineer based in Kochi, Kerala, holding a Bachelor of Computer Applications (BCA) from Mahatma Gandhi University. My work focuses on taking experimental GenAI prototypes and transforming them into robust, low-latency, containerized production systems.

At WebCastle Technologies, I architected bilingual (English & Malayalam) conversational AI agents using the Agno framework and persistent semantic memory with PostgreSQL pgvector (HNSW). My experience spans the complete data and backend lifecycle—from supervised ML modeling with Scikit-learn and Pandas at Zoople Technologies, to enterprise web operations at WellMade Network.

Currently available for full-time on-site, hybrid, or remote engineering roles in AI Engineering, Generative AI & RAG, Data Science, and Backend Systems.

Get in Touch
01 · Systems
6+
AI & Backend Systems
Production RAG, ML classifiers, and containerized APIs.
02 · Linguistic
Bilingual
English & Malayalam
Code-mixed tokenization, dialect sensitivity & prompt design.
03 · Latency
< 240ms
Retrieval Latency
pgvector HNSW indexing with asynchronous FastAPI coroutines.
04 · Availability
Immediate
Full-Time Ready
Open to on-site, hybrid (Kochi/BLR/India), or remote roles.
Featured Systems

Featured Production Systems

Real-world systems engineered for low latency, bilingual linguistic accuracy, and robust data integrity. Complete with open-source code and architectural documentation.

System 01GenAI & RAG
GitHub Repository
cogneva_system.production
Live
Cogneva — Privacy-First AI Therapist Platform Preview
Screens:
< 240ms
Retrieval Latency
English & Manglish
Language Support
94% Empathy Index
Evaluation Score
Dockerized
Deployment

Cogneva — Privacy-First AI Therapist Platform

Bilingual mental health conversational agent with long-term semantic memory and LLM-as-a-Judge validation.

The Engineering Challenge:Traditional digital mental health chatbots often lack clinical grounding, fail to remember past user disclosures across sessions, and are exclusively trained in standard English, alienating users who express emotional distress in regional languages like Malayalam.
Engineered Architecture & Solution:Built a hybrid RAG pipeline using Google Gemini and Agno framework coupled with pgvector for persistent conversational memory and semantic document retrieval. Implemented automated LLM-as-a-Judge evaluation scripts to measure empathy, safety guardrails, and contextual accuracy.
System Architecture Highlights:
FastAPI asynchronous backend with token-bucket rate limiting and session encryption
Agno agent orchestration with dynamic tool execution and emotional state tracking
PostgreSQL + pgvector vector store for cosine similarity chunk lookup
Stack:PythonFastAPIAgnoPostgreSQLpgvectorDockerGemini
System 02GenAI & RAG
GitHub Repository
documind_system.production
Live
DocuMind — Privacy-First Local PDF Intelligence Preview
Screens:
0% (100% Local)
Data Egress
100+ pages / 8s
Indexing Speed
< 45ms FAISS
Vector Search
$0 API Costs
Cost

DocuMind — Privacy-First Local PDF Intelligence

Local LLM document intelligence platform powered by RAG, Ollama, FAISS, and Streamlit.

The Engineering Challenge:Users handling proprietary legal contracts, internal financial reports, or personal academic papers cannot safely upload sensitive documents to public cloud AI platforms due to strict privacy and compliance risks.
Engineered Architecture & Solution:Designed a fully local RAG pipeline pairing Ollama-managed local open-weights LLMs with FAISS vector indexing and sentence-transformers embeddings. Engineered an interactive Streamlit UI for instant document Q&A and knowledge testing.
System Architecture Highlights:
Document ingestion engine with recursive character chunking and metadata preservation
Local vector database powered by FAISS for sub-50ms vector similarity lookup
Locally hosted LLM inference via Ollama ensuring zero external network egress
Stack:PythonStreamlitOllamaFAISSSentence-TransformersRAG
System 03Machine Learning
GitHub Repository
fake-invoice_system.production
Live
Fake Invoice Detection — ML Classification Preview
Screens:
96.4%
Model Accuracy
94.8%
Precision Score
< 15ms
Inference Latency
Random Forest
Algorithm

Fake Invoice Detection — ML Classification

Machine Learning fraud detection system utilizing Random Forest and interactive Streamlit interface.

The Engineering Challenge:Businesses suffer massive annual financial leakages due to fake invoices, inflated billing line items, and duplicate payment requests that easily bypass standard rule-based accounting filters.
Engineered Architecture & Solution:Engineered a multi-feature fraud classification system using Scikit-learn's Random Forest classifier. Extracted transactional ratios, vendor entropy, and billing anomaly indicators to predict fraud probability with high precision and low false-positive rates.
System Architecture Highlights:
Rigorous exploratory data analysis (EDA) and data wrangling on transaction datasets
Feature engineering: tax discrepancy scores, amount z-scores, vendor frequency
Random Forest ensemble model optimized with stratified cross-validation
Stack:PythonScikit-learnPandasRandom ForestStreamlitEDA

Additional Production Systems

Explore specialized AI evaluation, enterprise security, and transactional web architectures.

Technical Skills

Technical Competencies

A comprehensive, verified technical toolkit spanning modern AI frameworks, scalable asynchronous backends, and rigorous machine learning pipelines.

01 · Domain

Generative AI & LLMs

Production RAG, bilingual conversational agents, and semantic vector retrieval.

Gemini API & LLM Interactions
Agno Agent Framework
pgvector (HNSW Indexing)
FAISS Vector Search
Ollama Local Inference
Prompt Engineering
LLM-as-a-Judge Evaluation
Semantic Chunking & Embeddings
02 · Domain

Backend & Systems

Asynchronous microservices, connection pooling, and containerized deployment.

Python 3.11+ (Asyncio)
FastAPI Framework
Docker & Docker Compose
PostgreSQL & MySQL
Pydantic v2 Schemas
SQLAlchemy ORM
RESTful API Design
Linux Server Environment
03 · Domain

Data Science & ML

Supervised classification, feature engineering, and exploratory data analysis.

Scikit-learn Algorithms
Pandas & NumPy
Random Forest Classifiers
Feature Engineering
Anomaly Detection
Exploratory Data Analysis (EDA)
Power BI Dashboards
Model Evaluation Metrics
04 · Domain

Web & Operations

Full-stack web applications, CMS portal management, and technical digital operations.

Next.js & React 19
TypeScript & Tailwind CSS
PHP & MySQL Systems
WordPress Administration
Technical SEO & Auditing
Uptime Monitoring
Git & GitHub Workflows
Enterprise Portal Ops
End-to-End System Architecture Blueprint
Methodology: Bilingual RAG & Async FastAPI
Step 01 · Ingestion
Mixed Unicode Tokenizer

Code-mixed English & Malayalam input parsing with 768-dim dense vector projection.

Step 02 · Retrieval
pgvector HNSW Search

Sub-240ms semantic retrieval across indexed psychological and domain literature.

Step 03 · Reasoning
Agno Agent + Guardrails

Automated LLM-as-a-Judge validation for zero hallucination and high empathy scores.

Step 04 · Delivery
FastAPI Container Stream

Dockerized asynchronous SSE token streaming with zero client buffering under load.

Experience & Qualifications

Experience & Qualifications

Verified professional engineering history, academic computer science foundation, and accredited industry credentials.

Professional Experience3 Roles

Data Science Intern

WebCastle Technologies
Jan 2026 – Jun 2026Kochi, Kerala, India

Engineered production-grade conversational AI systems, semantic RAG architectures, and containerized backend services for enterprise clients.

Core Engineering Deliverables & Impact:
Built bilingual English & Malayalam conversational AI agent using Google Gemini and the Agno framework, enabling natural colloquial interactions.
Architected PostgreSQL + pgvector semantic RAG pipeline with optimized vector indexing for low-latency context retrieval.
Containerized modular FastAPI backend microservices using Docker and automated LLM-as-a-Judge validation scripts to benchmark response quality.
Collaborated closely with cross-functional engineering teams to integrate AI endpoints into responsive client interfaces.
Technologies:PythonFastAPIAgnoPostgreSQLpgvectorDockerGeminiRAG

Data Science Trainee

Zoople Technologies
Aug 2025 – Jan 2026Kerala, India

Completed intensive hands-on training spanning the full data science lifecycle, machine learning algorithms, and modern Generative AI workflows.

Core Engineering Deliverables & Impact:
Executed end-to-end data science projects: exploratory data analysis (EDA), data cleaning, feature engineering, and predictive modeling.
Developed supervised machine learning classification models using Scikit-learn and Pandas with cross-validation and hyperparameter tuning.
Constructed interactive data visualization dashboards using Power BI and Streamlit.
Implemented practical RAG pipelines and vector embedding workflows for document intelligence.
Technologies:PythonScikit-learnPandasPower BIStreamlitMachine LearningEDA

Technical Associate

WellMade Network / WellMade Network FZC LLC
Mar 2024 – Jul 2025Remote / UAE / India

Managed WordPress website infrastructure, news portal operations (gccbusinessnews.com), and technical SEO workflows.

Core Engineering Deliverables & Impact:
Oversaw ongoing website development, maintenance, and hosting configurations for enterprise news portal gccbusinessnews.com.
Implemented technical SEO audits, site speed optimizations, schema markup, and content indexing enhancements.
Managed digital content publishing workflows, social media automation, and operational uptime monitoring.
Assisted with PHP script customizations, plugin integrations, and database backups.
Technologies:WordPressTechnical SEOPHPgccbusinessnews.comDigital OperationsMySQL
Academic FoundationMG University

Bachelor of Computer Applications (BCA)

2022 – 2025 (Graduated Mar 2025)
Mahatma Gandhi (MG) University, Kerala

Specialization in Software Engineering, Data Structures, Relational Database Systems, and Object-Oriented Programming.

Higher Secondary Examination (Class XII)

2020 – 2022 (Completed Mar 2022)
Mary Matha HSS, Thrikkakara, Kochi

Science & Computer stream with foundational training in programming and mathematics.

Full-time collegiate computer science curriculum.
Accredited CertificationsVerified
Data Science Internship Certification
WebCastle Technologies, Kochi • 2026
Verified
Data Science & Generative AI Professional Trainee
Zoople Technologies • 2025
Verified
PHP & Python Web Development Certification
Riss Technologies • 2024
Verified
Verified Competency Standards
Bilingual Conversational RAG
pgvector HNSW Cosine Search
Docker Containerized APIs
Scikit-learn ML Pipelines
Contact

Get in Touch

Actively open to full-time engineering roles in AI, Generative AI (RAG & Agents), Data Science, and Backend Architecture.

Direct ChannelsAvailable
Immediate Joining Available

Open to full-time on-site, hybrid (Kochi/Bangalore/India), or remote engineering opportunities.

Direct Email
GetAnhar@gmail.com
Phone Number
+91-6238771132
Location & Timezone
Kochi / Ernakulam, Kerala, India (IST UTC+5:30)
Languages: EN, ML, HI, TA24h response time
Send a MessageDirect Message
Verified Dossier•Immediate Joining / Full-Time Ready

Anhar Eswaramangalam

AI & Data Science Engineer | Generative AI & Backend Systems

Immediate Joining
Executive Engineering Profile

Bachelor of Computer Applications (BCA) graduate specializing in production-ready bilingual Generative AI (RAG), asynchronous backend microservices, and machine learning pipelines. Experienced in designing vector chunk retrieval engines with pgvector (HNSW) and FAISS, building bilingual conversational agents with Google Gemini and the Agno framework, and deploying containerized FastAPI microservices with Docker.

Professional Industry Experience3 Roles

Data Science Intern

WebCastle Technologies
Jan 2026 – Jun 2026Kochi, Kerala, India

Engineered production-grade conversational AI systems, semantic RAG architectures, and containerized backend services for enterprise clients.

Built bilingual English & Malayalam conversational AI agent using Google Gemini and the Agno framework, enabling natural colloquial interactions.
Architected PostgreSQL + pgvector semantic RAG pipeline with optimized vector indexing for low-latency context retrieval.
Containerized modular FastAPI backend microservices using Docker and automated LLM-as-a-Judge validation scripts to benchmark response quality.
Collaborated closely with cross-functional engineering teams to integrate AI endpoints into responsive client interfaces.
PythonFastAPIAgnoPostgreSQLpgvectorDockerGeminiRAG

Data Science Trainee

Zoople Technologies
Aug 2025 – Jan 2026Kerala, India

Completed intensive hands-on training spanning the full data science lifecycle, machine learning algorithms, and modern Generative AI workflows.

Executed end-to-end data science projects: exploratory data analysis (EDA), data cleaning, feature engineering, and predictive modeling.
Developed supervised machine learning classification models using Scikit-learn and Pandas with cross-validation and hyperparameter tuning.
Constructed interactive data visualization dashboards using Power BI and Streamlit.
Implemented practical RAG pipelines and vector embedding workflows for document intelligence.
PythonScikit-learnPandasPower BIStreamlitMachine LearningEDA

Technical Associate

WellMade Network / WellMade Network FZC LLC
Mar 2024 – Jul 2025Remote / UAE / India

Managed WordPress website infrastructure, news portal operations (gccbusinessnews.com), and technical SEO workflows.

Oversaw ongoing website development, maintenance, and hosting configurations for enterprise news portal gccbusinessnews.com.
Implemented technical SEO audits, site speed optimizations, schema markup, and content indexing enhancements.
Managed digital content publishing workflows, social media automation, and operational uptime monitoring.
Assisted with PHP script customizations, plugin integrations, and database backups.
WordPressTechnical SEOPHPgccbusinessnews.comDigital OperationsMySQL
Core Technical Competencies
Generative AI & LLMs
Bilingual RAGGemini APIAgno Frameworkpgvector (HNSW)FAISSOllamaLLM Evals
Backend & Systems
Python 3.11+FastAPIDockerPostgreSQLMySQLPydantic v2Asyncio APIs
Data Science & ML
Scikit-learnRandom ForestPandas & NumPyFeature EngineeringEDAPower BI
Web & Operations
Next.js & React 19PHP & MySQLWordPress CMSTechnical SEOGit & GitHub
Academic Education

Bachelor of Computer Applications (BCA)

2022 – 2025 (Graduated Mar 2025)
Mahatma Gandhi (MG) University, Kerala

Specialization in Software Engineering, Data Structures, Relational Database Systems, and Object-Oriented Programming.

Higher Secondary Examination (Class XII)

2020 – 2022 (Completed Mar 2022)
Mary Matha HSS, Thrikkakara, Kochi

Science & Computer stream with foundational training in programming and mathematics.

Full-time computer applications curriculum.
Accredited CertificationsVERIFIED
Data Science Internship Certification
WebCastle Technologies, Kochi
2026
Data Science & Generative AI Professional Trainee
Zoople Technologies
2025
PHP & Python Web Development Certification
Riss Technologies
2024
All certificates verified and issued.