Skip to content

SRIMAN

Sriman written in Tamil
Illustration of Sriman in a denim shirt and glasses, seen from behind, looking to the right
LET’S
TALK

I build LLM agents, automations & AI pipelines.

GTM & AI Engineer orchestrating LLM agents and n8n automations, from lead sourcing and ICP scoring to RAG, outreach, and CRM sync.

10K+

Companies analyzed for intent signals

95%

Recommendation relevance (NDCG)

1

IEEE paper published

01 — About

About me

GTM Engineer and CSE graduate (B.Tech, 2026) who turns messy go-to-market processes into automations that save time and generate qualified leads. I focus on production reliability: LLM fallback chains, rate-limit handling, and workflows that keep running without babysitting.

Best Designer award, CSE Department Magazine, Dr. M.G.R. Educational and Research Institute.

Chennai, India

02 — Tools

Tools I build with

The stack behind my automations, AI agents, and pipelines.

  • n8nn8n
  • ClaudeClaude
  • PythonPython
  • HubSpotHubSpot
  • ApifyApify
  • SupabaseSupabase
  • FirecrawlFirecrawl
  • GitHubGitHub
  • Tailwind CSSTailwind CSS
  • ExaExa
  • Microsoft AzureMicrosoft Azure
  • JavaScriptJavaScript
  • ApolloApollo
  • ChromaDBChromaDB
  • CanvaCanva
  • MySQLMySQL
  • TriggifyTriggify
  • OpenRouterOpenRouter
  • FastAPIFastAPI
  • Google SheetsGoogle Sheets
  • BouncifyBouncify
  • OllamaOllama
  • DjangoDjango
  • Google CloudGoogle Cloud
  • lemlistlemlist
  • NVIDIA NemotronNVIDIA Nemotron
  • VS CodeVS Code
  • MongoDBMongoDB
  • HTML5HTML5
  • NeonNeon
  • CSS3CSS3

03 — Case studies

Selected work

01

LinkedIn Engagement-to-Outreach Automation

  • n8n
  • Apify
  • Claude
  • Firecrawl
  • Apollo
  • Lemlist

~3,000 engagers narrowed to 247 deduplicated leads with AI decision-maker filtering, email enrichment, and bounce verification.

  • Removed duplicate people who both liked and commented before qualification.
  • AI-assisted decision-maker qualification for patient-care organizations, recording the decision and rationale.
  • Apollo work-email lookup with attempt markers to prevent duplicate enrichment charges and allow retries.
  • Company-researched, personalized emails from post context, pushed into Lemlist with Google Sheets as a stage-by-stage queue.
02

Company Intent Signal Finder

  • n8n
  • Claude
  • Exa Search

~10,000 companies analyzed, with intent signals found for ~30% (hiring, acquisitions, expansion, EHR changes).

  • Claude agent with three Exa search paths (LinkedIn, job boards, company websites) returning structured JSON.
  • JSON repair, role checks, and recency windows: 30 days for hiring, 90 days for other events.
  • Agent failures routed to an “ERROR - rerun this row” marker for easy recovery.
03

LinkedIn Social Post-to-Outreach Pipeline

  • n8n
  • Apify
  • Apollo
  • Bouncify
  • Lemlist

~800 post authors per monthly run narrowed to 46 deduplicated leads.

  • AI decision-maker filtering plus Apollo email enrichment with Bouncify verification.
  • Guessed or extrapolated emails marked as not found instead of being validated.
  • Stage-level completion markers and unique-key row updates prevent duplicate processing and repeat campaign pushes.
04

Multi-Dimensional Organization Intelligence Pipeline

  • n8n
  • Apify
  • Firecrawl
  • LLM agents

Profiles companies across tech stack, buying signals, competitors, and leadership.

  • Parallel branches infer a company’s tools from LinkedIn job posts and its website, with Firecrawl as scrape fallback.
  • Industry-specific competitor branches (SaaS, insurance) with rate-limited LLM calls for stable batch runs.
  • A merge/dedup agent writes consolidated findings back to Google Sheets.
05

AI-Powered GTM & Lead Generation Pipeline

  • n8n
  • NVIDIA Nemotron
  • HubSpot

ICP scoring, research, personalized emails, reply detection, and primary/fallback LLM chains.

  • Sources and qualifies leads, scores them against an ICP, and researches each company before writing outreach.
  • Automated reply detection with AI-generated follow-ups, syncing lead status to HubSpot via SQL-based segmentation.
  • Primary/fallback LLM chain at every AI step keeps the pipeline running through model failures.
06

AI-Powered MERN Stack E-Commerce Site (IEEE TQCEBT'26)

  • MERN
  • DistilBERT
  • TF-IDF
  • scikit-learn

95% recommendation relevance (NDCG). Published on IEEE Xplore.

  • Multi-model NLP architecture (DistilBERT + TF-IDF + classical ML) to close a search-relevance gap in product discovery.
  • Research Lead & Builder; presented at IEEE TQCEBT'26, CHRIST University, Pune.
Read the paper

04 — Services

What I do

05 — Experience

Experience

  1. GTM Engineer Intern

    Aug 2026 – Oct 2026

    Healthbound Digital logoHealthbound Digital

    • Built a post-engagement outreach workflow: scrape likes and comments, enrich engagers, a Claude agent classifies decision makers, and personalized emails are pushed to lemlist.
    • Automated daily EOD reporting with a Telegram approval step before sending.

  2. Team Lead & Recruitment Coordinator

    2024 – 2025

    CSE Department Magazine logoCSE Department Magazine

    • Led 7+ contributors across the 38th and 39th editions.
    • Oversaw and designed each edition from text draft to polished, print-ready book.

06 — Toolkit

Skills & certs

GTM & Automation

  • n8n
  • Apify
  • Firecrawl
  • lemlist
  • ICP scoring

AI / GenAI

  • Claude
  • NVIDIA Nemotron
  • OpenRouter
  • RAG
  • Ollama

CRM & Tools

  • HubSpot
  • Google Sheets
  • Telegram

Backend

  • Python
  • Django
  • FastAPI

Databases

  • MySQL
  • MongoDB
  • Supabase
  • ChromaDB

Certifications

  • Anthropic – AI Fluency & Foundations (via CodePath)
  • AWS – Foundations of Prompt Engineering
  • Siemens Mobility – Project Management Job Simulation (Forage)
  • Cisco – Cybersecurity Essentials

06 — Contact

Let’s talk

Have a messy GTM process or a workflow that needs AI? Send me a note.

srimans2004@gmail.com
© 2026 Sriman SChennai, India