Abhigyan S.

I am a Solution Architect, Senior Software Developer, and AI Engineer with 20+ years of hands-on experience designing, building, and modernizing complex software systems. I combine architectural thinking with deep implementation expertise, delivering solutions that are practical, secure, scalable, and maintainable in real-world, business-critical environments. My background spans the Microsoft ecosystem—Windows desktop, web and API platforms, cloud-native systems, and data-intensive applications—across enterprise, e-commerce, CRM, and ERP domains. In parallel, I design end-to-end AI systems using NLP, LLMs, and RAG using ChatGPT, Gemini, Gemma, llama etc, integrating vector and Graph databases, semantic and hybrid search, and LangChain/LangGraph orchestration, Observability using OpenTelemetry while addressing cost, latency, reliability, and OWASP-aligned LLM security. I am an expert in preventing hallucinations that originate from data, training and inference, resulting in a grounded RAG application. The RAG applications also have guardrails to ensure that proper responses are presented to the end users. Architecturally, I work with Clean, Layered, Microservices, Event-Driven, Pipeline, and API-first designs, selecting patterns based on business constraints rather than trends. I also guide LLM selection through structured benchmarking—evaluating accuracy, hallucination risk, latency, and cost using frameworks such as LM-Eval-Harness, HELM, OpenAI Evals, and LangSmith—ensuring production-ready, data-driven decisions. Technical Expertise Natural Language Processing (NLP) • Text preprocessing: tokenization, stemming, lemmatization, n-grams • NLP using NLTK • Part-of-Speech (POS) tagging • Named Entity Recognition (NER) • Sentiment analysis • Text vectorization: Bag of Words, TF-IDF • Topic modeling: Latent Dirichlet Allocation (LDA), Latent Semantic Analysis (LSA) • Custom text classifiers using Logistic Regression, Naïve Bayes, Linear SVM Large Language Models (LLMs) • OpenAI models (ChatGPT APIs) • Transformer-based models: BERT, RoBERTa, DistilBERT, XLNet, clinicalbert • Local and self-hosted LLMs using Ollama and LMStudio - Gemma, MedGemma, Llama etc. AI Frameworks & Orchestration LangChain • Prompt engineering and prompt optimization • Output parsers and structured responses • LangChain Expression Language (LCEL) • Retrieval-Augmented Generation (RAG) pipelines • Similarity search, semantic search, and Maximal Marginal Relevance (MMR) search LangGraph • Stateful AI workflows • Short-term memory (STM) and long-term memory (LTM) design • Multi-step and agent-based execution graphs Observability: • LangSmith Vector Databases & Storage • Pinecone, ChromaDB, Qdrant, Turso, SQLite, SQL Server 2025, MongoDb, Neo4j, postgresql • Vector and hybrid search architectures for RAG systems LLM Application Security (OWASP-Aligned) • Prompt injection prevention • Protection against sensitive information disclosure (PII, PHI) (Microsoft Presidio) • Supply chain vulnerability mitigation • Data poisoning and model poisoning defenses • Secure output handling and validation • Managing excessive agency and tool misuse • System prompt leakage prevention • Vector and embedding security weaknesses • Misinformation and hallucination risk mitigation • Unbounded consumption and cost control • Model denial-of-service (DoS) mitigation • Secure plugin and tool design • Reducing overreliance on LLM outputs • Model theft prevention strategies • LLM security best practices and threat modeling Policy Engines: • Open Policy Agent (OPA) • Casbin • Cedar Microsoft & .NET Development • VBA, VB 6.0 • .NET Framework, ASP.NET Core, ASP.NET MVC • REST APIs, Web APIs, Minimal APIs, Microservices • WPF, WinUI 3, .NET MAUI • Windows Desktop and Enterprise Application Development Cloud & Backend Systems • Azure Development • API-driven and service-oriented systems Databases & Performance Engineering • SQL Server administration, configuration, and performance tuning • Query optimization, indexing strategies, execution plan analysis • SQL Server Cluster Instances (installation and management) • Dapper, NHibernate, PetaPoco Office & Automation • Microsoft Office Add-ins • Office 365 Add-ins • VSTO Add-ins • Advanced VBA Macros Legacy & Specialized Systems • Visual FoxPro • Classic ASP • Btrieve-based applications What I Can Help You With NLP pipelines and custom text classifiers LLM-powered applications and chatbots RAG systems using Pinecone, ChromaDB, Qdrant, SQLServer, SQLite etc. Secure LangChain / LangGraph workflows Semantic and hybrid search over structured and unstructured data Speech-to-text solutions and evaluation LLM security reviews and hardening I look forward to working with you and I firmly believe in long and mutually beneficial relationship with my clients. Thank you for taking time to read this. Regards, Abhi