# Swarnava Dutta — Lead AI Architect > Swarnava Dutta is a Lead AI Architect and enterprise AI consultant based in Kolkata, India, with 6 years of generative AI engineering experience. He architects and ships production-grade agentic AI systems, multi-agent frameworks (LangGraph, MCP), enterprise-scale RAG pipelines, and LLM fine-tuning (LoRA/PEFT) with RAGAS-driven evaluation — systems that have scaled to 10K+ daily queries at 99.9% uptime and cut manual workloads by up to 67%. Currently Senior GenAI Engineer and Technical Lead at TEKsystems Global Services. Available for full-time roles, contract projects, and freelance consulting; experienced with distributed US and EU teams. Website: https://swarnava.dev | Profiles: https://github.com/swarnava-dutta, https://linkedin.com/in/swarnava-dutta, https://x.com/Swarnava_Duttaa # About Swarnava Dutta Swarnava Dutta is a Lead AI Architect and enterprise AI consultant based in Kolkata, India, with 6 years of generative AI engineering experience. He architects and ships production-grade agentic AI systems, multi-agent frameworks (LangGraph, MCP), enterprise-scale RAG pipelines, and LLM fine-tuning (LoRA/PEFT) with RAGAS-driven evaluation — systems that have scaled to 10K+ daily queries at 99.9% uptime and cut manual workloads by up to 67%. Currently Senior GenAI Engineer and Technical Lead at TEKsystems Global Services. Available for full-time roles, contract projects, and freelance consulting; experienced with distributed US and EU teams. ## Focus I design agentic workflows, production RAG pipelines, and LLMOps infrastructure enterprise AI systems built to hold up under real traffic, real users, and real compliance pressure. Core expertise: Multi-Agent Systems, Fine-tuning & LoRA, RAG Architecture, MCP Workflows. Topics: Generative AI, Large Language Models, LangGraph, LangChain, RAG Architecture, AI Agents, Multi-Agent Systems, Model Context Protocol (MCP), Azure OpenAI, Azure AI Foundry, Prompt Engineering, LLM Fine-tuning and LoRA, LLMOps, Vector Databases, Enterprise AI Systems. ## Education Bachelor of Technology in Computer Science and Engineering, Maulana Abul Kalam Azad University of Technology (2016-2020). ## Availability Status: Available for New Opportunities. Open to: Full-time Opportunities, Contract Projects, Freelance Consulting. Work setup: Remote & hybrid friendly. Response time: Within 12 hours. # Services offered by Swarnava Dutta > AI consulting and delivery services across multi-agent systems, RAG architecture, model fine-tuning, and enterprise AI platforms. ## Multi-Agent Systems Design multi-agent systems that coordinate tools, memory, approvals, and specialist roles without becoming hard to manage. - Best suited for: Teams replacing repetitive manual work with planner, reviewer, and executor workflows. - Typical outcomes: Clear agent roles and handoffs; Human review steps and safety rails; Monitoring points for improvement over time - Stack: LangGraph, MCP, Tool calling, Guardrails ## RAG Architecture & Vector Search Build retrieval systems that surface the right context quickly, keep answers grounded, and make knowledge easier to use. - Best suited for: Products and internal copilots that need trusted answers from documents, data systems, and live business context. - Typical outcomes: Ingestion, chunking, and indexing plan; Hybrid retrieval with reranking and citations; Evaluation loops for relevance and answer quality - Stack: Azure AI Search, Pinecone, Reranking, RAG evals ## Model Fine-tuning & Optimization Improve quality, speed, and consistency with the right mix of fine-tuning, prompt shaping, dataset design, and inference optimization. - Best suited for: Teams that already see promise but need better domain accuracy, tone control, or task fit. - Typical outcomes: Dataset curation and benchmark framing; LoRA or PEFT fine-tuning workflow design; Latency, cost, and quality optimization paths - Stack: LoRA, PEFT, Benchmarks, Inference tuning ## Enterprise AI Architecture Shape the end-to-end AI setup so it is secure, observable, and easy for real teams to run over time. - Best suited for: Teams moving from early AI experiments to dependable platforms, governance, and long-term use. - Typical outcomes: A reference setup for services and data flow; Security, compliance, and deployment patterns; Monitoring, team handoff, and maintainability - Stack: Azure, Kubernetes, CI/CD, Monitoring ## Engagement models Baseline rates and delivery estimates are shared during the first scoping call. - **AI Architecture Audit** (1-2 weeks · Fixed scope): A structured review of an existing AI system: retrieval quality, agent design, evaluation coverage, cost, and reliability risks. Deliverable: Scored findings report with a prioritized remediation roadmap. - **Discovery Sprint** (2-3 weeks · Fixed scope): Shape an AI idea into a validated plan: use-case framing, data readiness, architecture options, and a working proof of concept. Deliverable: Reference architecture, PoC, and a delivery estimate. - **End-to-End Delivery** (4-12 weeks · Milestone-based): Design and ship a production AI system: agent workflows, RAG pipelines, evaluation loops, deployment, and team handoff. Deliverable: Production system with monitoring, documentation, and handoff. - **Fractional AI Architect** (Monthly retainer): Ongoing architecture ownership for teams that need senior AI leadership without a full-time hire: reviews, roadmaps, delivery oversight. Deliverable: Weekly architecture sessions plus async design reviews. # How Swarnava Dutta runs AI engagements > A five-step delivery process where every step produces something visible and testable. 1. **Discover & scope** — I start by understanding your workflow, your users, and your data — and what success would actually look like. No code until that's clear. Deliverables: Problem definition, Success metrics, Scope document. 2. **Design the architecture** — Then I pick the retrieval strategy, agent topology, models, and guardrails that fit your constraints — not whatever is trending that week. Deliverables: Architecture diagram, Model selection, Risk review. 3. **Build in thin slices** — I ship a working end-to-end slice early, then iterate against real data and real user feedback. You see progress every week, not at the end. Deliverables: Working prototype, Weekly demos, Iteration log. 4. **Evaluate & harden** — Before launch I build eval pipelines, tune latency and cost, and make sure failures show up on a dashboard — not in a user complaint. Deliverables: Eval suite, Latency budget, Observability. 5. **Launch & hand off** — I deploy with monitoring in place, write the runbooks, and make sure your team can run the system confidently without me. Deliverables: Production deploy, Runbooks, Team enablement. # Projects by Swarnava Dutta > Production AI systems: multi-agent workflows, RAG platforms, copilots, and automation with measurable outcomes. ## LindsAI — Multi-agent automation with JIRA, SharePoint & live web search Built a multi-agent system that turns untracked email intake into a single auditable, session-persistent channel. Specialized agents use tool calling into JIRA and SharePoint plus dynamic web search to gather context, ask clarifying questions, classify risk, and raise tickets automatically — with human-in-the-loop approval checkpoints before anything ships. What used to take teams 2+ days from intake to ticket now lands in under 10 minutes, with parallelized Redis-backed workers compressing P95 pipeline latency by 85%. - Technologies: LangGraph, JIRA & SharePoint Tools, Web Search, Redis - Highlights: 98% intent accuracy, <10 min intake to ticket ## GRIT — Adaptive sales enablement for regulated medical teams Built a persona-adaptive sales enablement platform for medical technology teams that need faster onboarding without losing regulatory context. The system combines RAG, MCP-style content updates, and skill-gap analysis so each rep can practice against role-specific objections, product narratives, and compliant messaging. It turns static enablement material into an interactive coaching workflow that helps managers see where reps are stuck and gives new hires a clearer path from training to confident field conversations. - Technologies: RAG, MCP, LangGraph, Real-time AI - Highlights: 65% faster onboarding, 50K+ monthly sessions ## Talksmith — Voice-first analytics for teams that live in data Created a voice-enabled data copilot that lets analysts and business users ask questions naturally instead of writing SQL by hand. The workflow captures spoken intent, translates it into structured database queries, and returns usable answers from Snowflake-backed data sources. LangGraph coordinates the voice, reasoning, SQL generation, and response steps so users can move from a business question to a data-backed answer in seconds while keeping the interaction simple enough for non-technical teams. - Technologies: LangGraph, Snowflake, Voice AI, SQL Gen - Highlights: 200+ daily users, 8 business teams ## IntervueRecall — Interview recordings turned into clean question lists Built IntervueRecall as a SaaS workspace for candidates who want a clear record of the questions asked in their interviews. Users upload an interview recording, the app transcribes the audio, extracts only the interviewer questions, and presents them as a clean list. The workflow is focused on question recall so candidates can review what was asked and prepare for future interviews with a more accurate question bank. - Technologies: Audio Transcription, Question Extraction, Interview AI, SaaS - Highlights: 1-click audio upload, Qs extracted list ## InsightDesk AI — A knowledge copilot for documents, tables, and daily decisions Built an enterprise assistant that can decide whether a question should be answered from documents, structured tables, or a combination of both. The system uses Databricks, RAG, Delta-backed data, and tool-style routing so employees can ask operational questions without knowing where the answer lives. It is designed for repeated internal use: clear answers, grounded context, and a workflow that reduces the manual analyst effort normally spent searching files, checking dashboards, and stitching together evidence. - Technologies: Databricks, RAG, Delta Tables, MCP - Highlights: 10K+ monthly questions, ~50% analyst lift ## LinkedInfluencer — AI content intelligence for building a credible LinkedIn voice Built LinkedInfluencer, a local browser app that discovers live GenAI discussions from Reddit, Hacker News, Dev.to, and Stack Overflow, scores story relevance, and turns the strongest ideas into LinkedIn-ready posts. The LangGraph pipeline runs discovery, aggregation, analysis, selection, hook generation, content transformation, sanitization, quality control, and final JSON export with live server-sent progress events in the UI. It supports Fast and Deep modes, configurable audience, goal, tone, source selection, worker limits, output folders, and Claude token budgets so content can be tuned without touching code. - Technologies: LangGraph, Content AI, A/B Testing, Automation - Highlights: 20+ hrs saved weekly, A/B content optimization ## FitScout — Recruiter intelligence for faster resume screening Built FitScout as a recruiter-facing screening assistant that turns a pile of resumes into structured hiring signals. It uses LlamaIndex and Pinecone to extract skills, experience, role fit, and searchable candidate context, giving recruiters a faster way to compare applicants without reading every document from scratch. The product focuses on practical hiring workflows: resume ingestion, skill matching, evidence-backed summaries, and shortlists that still leave the recruiter in control of the final judgment. - Technologies: LlamaIndex, Pinecone, FastAPI, OpenAI - Highlights: 1000+ resumes monthly, 90%+ screening accuracy ## WhisperIt Legal Copilot — A private document assistant for faster legal review workflows Designed a polished legal AI assistant for Swiss document workflows, focused on making contract review, clause summarization, and legal Q&A feel fast, controlled, and professional. The experience presents a clean chatbot interface over a privacy-conscious RAG flow, helping users move from dense legal text to practical answers, draft language, and review-ready summaries. The project emphasizes trust signals that matter in legal operations: source-grounded responses, clear document context, GDPR-aware handling, and a calmer workflow for teams that need accuracy without slowing down routine review. - Technologies: RAG, GDPR, Legal Tech, Semantic Search - Highlights: 45% faster turnaround, GDPR privacy aligned ## OpsAnswer Hub — One assistant for HR, finance, policy, and payroll questions Integrated HR and finance knowledge into a shared AI assistant for routine internal support. The chatbot answers policy questions, surfaces payroll information, explains process details, and routes users toward the right operational context without forcing them through multiple systems. Built with LangChain, Azure OpenAI, Databricks, and ETL pipelines, it is shaped around the kind of questions employees ask repeatedly and the kind of auditability support teams need when answers involve people, money, and policy. - Technologies: LangChain, Azure OpenAI, Databricks, ETL - Highlights: 40% routine query share, 200+ hrs saved monthly # Professional experience of Swarnava Dutta > 6 years across generative AI engineering, cloud data platforms, and production AI delivery. ## Senior Generative AI Engineer | Technical Lead — TEKsystems Global Services Pvt. Ltd Dec 2025 - Present - Cut intake-to-ticket turnaround from 2+ days to under 10 minutes, compressing P95 pipeline latency by 85% with parallelized Redis-backed workers - Improved learner skill-gap detection accuracy by 37% across 6K+ monthly sessions with adaptive quiz generation and continuous RAGAS evaluation Tech: LangGraph, MCP, Azure AI Foundry, Azure AI Search, Redis, RAGAS, FastAPI ## Senior Generative AI Engineer | Technical Lead — Syren Technologies Private Limited Jul 2025 - Dec 2025 - Scaled a supply chain AI copilot to 10K+ daily queries, reducing latency by 63% while maintaining 99.9% uptime - Boosted contextual recall by 34%, improved multi-turn response accuracy by 28%, and reduced context errors by 42% Tech: LangGraph, Pinecone, FastAPI, Mistral-7B, LoRA, MCP, BLEU/ROUGE-L ## Senior Generative AI Engineer | Technical Lead — RWS Moravia India Private Limited Jan 2024 - Jul 2025 - Launched the company's first production-grade multilingual enterprise RAG platform and AI agent - Cut manual knowledge lookup by 43% and established rigorous BLEU and NDCG evaluation benchmarks Tech: RAG, Multilingual AI, LangChain, Azure OpenAI, BLEU, NDCG ## Junior Data Scientist — BinaryERP Software Private Limited Jul 2022 - Dec 2023 - Built semantic search prototypes that drove adoption across 13 B2B users - Improved dataset usability by 26% and laid the foundation for enterprise NLP initiatives Tech: Python, NLP, Embeddings, Text Classification, Semantic Search, Data Pipelines ## Associate Data Engineer — Capgemini Technology Services India Limited Jan 2021 - Jul 2022 - Reduced nightly batch runtime by 95% for a global banking client - Decreased batch failures by 61% while preserving SLA compliance and improving ETL validation automation Tech: Azure Data Lake, Azure Data Factory, SQL Server, ETL, Data Validation, Automation # Skills of Swarnava Dutta > Technical toolbox across GenAI, Azure AI, evaluation, vector search, LLMOps, and data engineering. ## GenAI & LLMs Building end-to-end LLM systems, agent workflows, and prompt strategies teams can rely on. RAG Architecture, Agentic AI, Model Context Protocol, LangGraph, LangChain, LlamaIndex, Hugging Face, OpenAI, Azure OpenAI, LoRA/PEFT, Prompt Engineering, BERT/T5, AI Agents, A2A Protocol, Human-in-the-Loop ## Azure AI Building AI solutions on Azure with cloud services, search, deployment patterns, and data tooling. Azure AI Foundry, Copilot Studio, Azure AI Search, Azure Functions, Azure Key Vault, PromptFlow, Azure Kubernetes, Azure Data Lake, Azure Databricks, Azure Data Factory ## Evaluation & Quality Measuring retrieval and answer quality so AI systems improve with clear feedback loops. BLEU, ROUGE, Recall@k, NDCG, F1 Score, RAGAS, LangSmith, Langfuse, Promptfoo ## Vector Databases Building retrieval layers with semantic indexing, hybrid search, and embedding pipelines for fast, accurate lookup. Pinecone, FAISS, ChromaDB, Pgvector, Hybrid Search, Semantic Search, Embeddings ## LLMOps Keeping AI applications dependable with deployment workflows, monitoring, orchestration, and automation. Git, Docker, Kubernetes, FastAPI, CI/CD, Airflow, Redis, MLflow, Azure DevOps, Monitoring ## Programming & Data Strong foundations in Python, ML tooling, and data platforms that support dependable delivery. Python, PyTorch, TensorFlow, SQL, Pandas, NumPy, PySpark, TypeScript/React, Kafka, Data Modeling, Unity Catalog, Delta Lake # Certifications of Swarnava Dutta > Cloud, data, and AI credentials from Microsoft, Databricks, and Apache. - **Azure Fundamentals** (AZ-900) — Microsoft. Core Azure services, security, pricing, and cloud basics. Verify: https://www.credly.com/badges/040e2d9c-5040-43c6-8bd4-112c15119e49/public_url - **Azure Data Fundamentals** (DP-900) — Microsoft. Azure data services, storage models, and analytics foundations. Verify: https://www.credly.com/badges/2e45cb7e-b3f5-4bd1-8239-8a8b88471f1a/public_url - **Databricks Fundamentals** — Databricks. Lakehouse concepts, workspace workflows, and platform essentials. Verify: https://credentials.databricks.com/3f5f5288-4ed7-402a-8985-afb5f86784cd#acc.5hvJUI0d - **Generative AI Fundamentals** — Databricks. LLM concepts, prompt basics, and responsible GenAI grounding. Verify: https://credentials.databricks.com/161817b5-4d1d-49bb-a199-a45b74dd2253 - **Azure Databricks Platform Architect** — Databricks. Secure, scalable Azure Databricks architecture and platform design. Verify: https://credentials.databricks.com/df3c73ea-4b50-4f5c-ae16-351172a5ff0e - **Apache Airflow Fundamentals** — Apache. DAG design, scheduling, and workflow orchestration fundamentals. Verify: https://www.credly.com/badges/2e45cb7e-b3f5-4bd1-8239-8a8b88471f1a/public_url ## Education Bachelor of Technology in Computer Science and Engineering, Maulana Abul Kalam Azad University of Technology (2016-2020). # Frequently asked questions about Swarnava Dutta ## Who is Swarnava Dutta? Swarnava Dutta is a Lead AI Architect based in Kolkata, India, with 6 years of experience in generative AI engineering. He designs and ships production-grade AI systems — multi-agent workflows built with LangGraph, retrieval-augmented generation (RAG) architectures, and enterprise LLM platforms on Azure OpenAI and Azure AI Foundry. His most recent project is LindsAI, a multi-agent system with tool calling into JIRA and SharePoint plus dynamic web search. He currently works at TEKsystems as a Senior GenAI Engineer and Technical Lead. ## What kind of work can I hire you for? Four things, mainly: designing multi-agent AI systems (LangGraph, MCP, tool calling into platforms like JIRA and SharePoint, guardrails), building RAG architecture and vector search that actually retrieves the right things (Azure AI Search, Pinecone, hybrid search, reranking, RAGAS evaluation), LLM fine-tuning and optimization (LoRA/PEFT, dataset design, inference tuning), and enterprise AI architecture that takes teams from pilot to a secure, observable production platform. If your problem doesn't fit one of those boxes neatly, ask anyway — most interesting problems don't. ## What's your day-to-day tech stack? LangGraph and LangChain for orchestration, Azure OpenAI and Azure AI Foundry for models, Pinecone and Azure AI Search for retrieval, Snowflake and Databricks on the data side, FastAPI for serving, and MCP for tool integration. I work across the full LLM lifecycle — prompting, RAG, fine-tuning with LoRA, agent orchestration, evaluation, and LLMOps. ## Are you available right now? Yes — I'm open to full-time roles, contract projects, and freelance consulting. The fastest way to reach me is booking a quick call at cal.com/swarnava-dutta/quick-call, or connect on LinkedIn, or email swarnava.dev@gmail.com. I usually reply within 12 hours. ## What does working with you look like? I keep it to five steps: first we get clear on the problem and what success measurably means, then I design the architecture — retrieval, agents, models, guardrails. I ship a working end-to-end slice early and demo every week, build evaluation and observability before launch, and hand off with runbooks so your team owns it without me. No long silent build phases, no big-bang launch surprises. ## Do you work with remote or international teams? Yes — most of my work has been with distributed teams across the US and Europe, so collaborating across time zones is the norm for me. I'm based in Kolkata, India (IST). Contracts, full-time remote roles, and hybrid arrangements within India all work. # Contact Swarnava Dutta > Available for New Opportunities. Response time: Within 12 hours. - Email: swarnava.dev@gmail.com - Schedule a 15-minute scoping call: https://cal.com/swarnava-dutta/quick-call - LinkedIn: https://linkedin.com/in/swarnava-dutta - GitHub: https://github.com/swarnava-dutta - X/Twitter: https://x.com/Swarnava_Duttaa - Location: Kolkata, India (IST) — Remote & hybrid friendly - Resume PDF: https://swarnava.dev/resume/Swarnava-Dutta-Lead-AI-Architect.pdf Open to: Full-time Opportunities, Contract Projects, Freelance Consulting.