Freelancers using LangChain in IndiaFreelancers using LangChain in India
ProjectsPeople
Anmol Baranwal's avatar
Anmol Baranwal
India
Technical Writer (1.5M+ reads) & Open Source Developer
$10k+
Earned
23x
Hired
56
Followers
BUILDSHIP_EXPERT

expert

HIRED_BY_YC_EXPERT

Hired

Technical Writer (1.5M+ reads) & Open Source Developer
Cover image for Technical Writing for Copilotkit | 40k+ views
Technical Writing for Copilotkit | 40k+ views
1
84
Cover image for Technical Writing for Encore | 30k+ views
Technical Writing for Encore | 30k+ views
1
17
Cover image for Technical Writing for Encore | 30k+ views
Technical Writing for Encore | 30k+ views
0
9
Cover image for Technical Writing for Opire | 60k+ Views
Technical Writing for Opire | 60k+ Views
0
20
LangChain(1)
Trashu Vashisth's avatar
Trashu Vashisth
Delhi, India
Building Production-Grade AI Agents & RAG Systems
13
Followers
Building Production-Grade AI Agents & RAG Systems
Cover image for The Problem: Sales teams waste 60%
The Problem: Sales teams waste 60% of their time researching leads instead of closing them. The Solution: I built a custom Agentic AI Pipeline that automates deep-dive business intelligence and lead scoring. Key Technical Highlights: Multi-Agent Architecture: Built using CrewAI, featuring a 'Business Intelligence Specialist' (for real-time research) and a 'Senior Sales Director' (for strategic scoring). High-Speed Intelligence: Powered by Llama 3.3-70B  for near-instant reasoning and decision-making. Real-time Web Scoping: Integrated Tavily AI to fetch live revenue data, employee counts, and market positioning. Enterprise Storage: A robust SQLite backend to manage lead pipelines with a sleek Streamlit dashboard. Smart Throttling: Engineered custom rate-limiting and token-trimming logic to ensure 99.9% uptime even under heavy API constraints. How it works: Simply enter a company name and URL. The AI agents scour the web, analyze the company's "AI potential," calculate a priority score (0-100), and even write a personalized sales pitch—all in under 30 seconds.
0
63
Cover image for Developed a production-grade Retrieval-Augmented Generation
Developed a production-grade Retrieval-Augmented Generation (RAG) system specifically designed to automate the analysis of complex Environmental, Social, and Governance (ESG) reports. This tool bridges the gap between static LLMs and the dynamic, data-heavy requirements of legal and sustainability compliance. [1 (https://www.youtube.com/watch?v=wkYPcMtwlN8)] Key Features & Capabilities Intelligent Document Processing: Automatically handles large, unstructured PDF/Word ESG reports, extracting critical clauses and metrics in seconds. Fact-Grounded Q&A: Uses a RAG architecture to ensure all answers are strictly based on the uploaded documents, virtually eliminating AI hallucinations. Compliance Mapping: Cross-references internal company data with global frameworks like CSRD, GRI, and TCFD to identify gaps or inconsistencies. Audit-Ready Traceability: Every insight generated includes direct citations and excerpts from the source files, providing a clear "paper trail" for legal teams. Automated Drafting: Capability to draft legal summaries, notices, or internal policy updates based on analyzed ESG risks Note: The 'Slaughter and May' branding in the sidebar is for UI/UX demonstration purposes only, showcasing how the tool integrates into a top-tier law firm's environment. #AI #RAG #LegalTech #ESG #Python #LangChain
1
151
Cover image for Developed a full-stack RAG-based E- Commerce
Developed a full-stack RAG-based E- Commerce AI chatbot using React.js and Tailwind CSS that suggests the perfect laptop from a live catalog. Integrated ChromaDB with BGE Embedding models to provide highly accurate, context-aware product recommendations and instant technical support." Key Highlights: Smart Laptop Recommendations: Uses Semantic Search to match user needs (gaming, coding, etc.) with real-time specs. Advanced Tech Stack: Powered by LangChain for orchestration and BGE models for superior data retrieval. Modern UI/UX: Built a responsive, clean interface using React.js and Tailwind CSS. Zero Hallucination: Ensures all suggestions are strictly grounded in the available product inventory.
2
1
241
Cover image for Built a highly scalable Retrieval-Augmented Generation
Built a highly scalable Retrieval-Augmented Generation (RAG) chatbot designed to interact with private datasets/PDFs. Unlike standard LLMs, this system minimizes hallucinations by retrieving real-time context from a local knowledge base before generating responses. Key Features: Semantic Search: Implemented Vector Embeddings to perform high-speed similarity searches across thousands of document chunks. Smart Retrieval: Integrated a retrieval pipeline using LangChain to fetch the most relevant context for user queries. Source Citation: Configured the bot to provide source references from documents, ensuring data transparency and accuracy. Optimized Performance: Used FAISS/Chromadb for efficient vector storage and retrieval.
4
295
LangChain(5)
Saket Panwar's avatar
Saket Panwar
proRohtak, India
AI Agents & LLM Systems · Ex-Amazon & Microsoft · 10+ Years
5.0
Rating
1
Followers
AI Agents & LLM Systems · Ex-Amazon & Microsoft · 10+ Years
Cover image for Healthcare AI-Based Decision Support System Architected
Healthcare AI-Based Decision Support System Architected a multi-step LLM-powered clinical decision support platform that transforms patient records and medical guidelines into structured diagnostic workflows. The system combines OCR processing, patient-data ingestion, AI-powered reasoning, and human-in-the-loop validation to improve reliability and consistency of clinical recommendations. Key contributions included: • Designing the LLM decision-engine architecture • Building medical guideline-to-decision-tree workflows • Developing evaluation and review tooling for quality assurance • Integrating OCR and patient-data processing pipelines • Supporting orchestration across multiple AI and healthcare system components Technologies: Python, LLMs, OCR, Workflow Orchestration, Healthcare AI
0
24
Cover image for SmartCart RAG : AI Support
SmartCart RAG : AI Support Assistant on Internal Knowledge Base Built a production RAG chat app over SmartCartCommerce's internal knowledge base - 86 documents, 856 indexed chunks covering policies, support playbooks, seller operations, and the brand admin console. Implemented three distinct AI personas (Customer, Concierge, Brand Partner) each with tailored retrieval and system prompts. Queries are embedded via OpenAI, semantically searched in Pinecone, and streamed back with source citations. Replaced manual doc-hunting for support teams with instant, grounded answers.
0
27
Cover image for AgentOS : Multi-Agent AI Orchestration
AgentOS : Multi-Agent AI Orchestration Platform Built AgentOS, a production-ready multi-agent orchestration platform that enables teams to deploy, monitor, and coordinate AI agents for complex workflows. Designed a modular architecture where specialized agents collaborate - handling tasks like research, code generation, and decision-making in parallel. Implemented real-time agent communication, persistent memory, tool-use pipelines, and a dashboard for observability. The system reduces manual intervention in repetitive workflows by over 80% and scales seamlessly across different LLM providers.
0
23
Cover image for Amazon Subscribe & Save —
Amazon Subscribe & Save — Perl → Java Migration
0
17
LangChain(3)
Anurag Nagare's avatar
Anurag Nagare
Mumbai, India
I’m an AI & Machine Learning engineer with expertise in deve
I’m an AI & Machine Learning engineer with expertise in deve
Cover image for Most AI research tools are
Most AI research tools are just a chatbot with a search button. I built something different. Every time you ask an AI to research something, you're getting one model, one pass, no quality check. It writes confidently, cites poorly, and you have no idea if what it produced is actually accurate. For anyone making real decisions from AI-generated research, that's a silent risk most people ignore. The problem gets worse at scale the longer and more complex the question, the more a single model hallucinates, misses sources, and loses structure. There's no one checking its work. So I built ResearchOS a 5-agent pipeline where each agent has one job. A Supervisor breaks down your question. A Researcher runs parallel searches across 22+ sources. An Analyst extracts data and auto-generates charts. A Writer synthesises a cited report. A Critic fact-checks it and sends it back for revision if anything is wrong. The loop runs up to 3 times before the report is approved. One question in. A full cited report with charts and PDF export in under 10 minutes. I tested it live by watching the Critic catch a missing citation mid-run and send the Writer back to fix it before approval. That's the part that makes this actually usable for real work. Built on LangGraph, Groq, Tavily, ChromaDB and runs entirely on free tiers.
1
78
Cover image for HybridAlpha (Hybrid RAG) : One tool
HybridAlpha (Hybrid RAG) : One tool digs into actual SEC filings, not just static documents. From EDGAR, it grabs 10, Ks and 10, Qs fresh each time. Sections like MD&A or Risk Factors get split out by name during parsing. Storage happens two ways at once: words go to ChromaDB, numbers land in SQLite. When a question arrives, the router decides, tone, driven, number, heavy, or both. Depending on that choice, the query moves to one place, sometimes both. Context flows forward only after sorting is done. Answers come from Llama 3.3 70B via Groq, always tagged with sources. Each output ties back to where the data lived. Start by asking, What risks did Apple highlight regarding AI rivals? Out comes exact quotes pulled straight from official documents.
0
49
Cover image for I created WealthWise Agent, a smart personal finance planner...
I created WealthWise Agent, a smart personal finance planner designed to craft personalized budget plans and investment strategies. This app takes into account user inputs like salary, expenses, and financial goals, and then uses a Large Language Model (Gemini) to analyze these factors based on the 50/30/20 budgeting rule. It offers a clear step-by-step reasoning log, a detailed JSON-structured financial plan, and an interactive visualization of budget allocation, empowering users to make informed choices to reach their financial goals. 💻 Tech Stack Used: Frontend/UI: Gradio (custom themed with CSS, Orbitron font) AI/Logic: Google Gemini (gemini-1.5-flash) with LangChain agents Data: yFinance API for real-time stock/ETF data, Pandas & NumPy for calculations Visualization: Plotly Express for interactive charts
0
51
Cover image for What your attention heatmap isn't
What your attention heatmap isn't telling you Everyone's staring at attention heatmaps and calling it "interpretability." Almost nobody's asking whether a single attention map actually tells you what the model used to make its decision. It doesn't. Not on its own. A raw attention map from layer 8 shows you what layer 8 attended to. It says nothing about how that signal got mixed, diluted, or overwritten by every layer before and after it. Attention rollout fixes this — and I built a walkthrough to show why it matters. Here's what makes it more than a "pretty heatmap" demo: Instead of visualizing one layer's attention, I traced how information actually flows through the full transformer stack. → Every layer's attention matrix is extracted, per head, per token → Multi-head attention is averaged, then combined with the residual connection (identity + attention) — this is the step most tutorials skip, and it's the one that actually matters → The combined matrices are matrix-multiplied layer by layer, rolling attention forward from input to output → The result: a single map showing genuine token-to-token influence across the entire network, not just one layer's snapshot The overlay shows you everything: → Per-layer attention vs. rolled-out attention, side by side → Token importance scores overlaid directly on the input text → A comparison view: which tokens raw attention says "matter" vs. which ones rollout says actually matter → Head-level breakdown so you can see which heads specialize vs. which are noise No black box. No "trust me, the model looked here." Just linear algebra, applied honestly across every layer instead of cherry-picking one. Built with PyTorch + HuggingFace Transformers + Matplotlib. Runs on any pretrained transformer, fully offline. ⚠️ Important: attention rollout is an approximation, not ground truth. It assumes attention is the primary information pathway, which ignores MLP layers and can still mislead for very deep models. Treat it as a debugging lens, not proof of causality.
0
26
LangChain(3)
Mehul Sethia | Senseibles's avatar
Mehul Sethia | Senseibles
proIndia
Design, Build, AI, Automation
$50k+
Earned
17x
Hired
5.0
Rating
61
Followers
Design, Build, AI, Automation
Cover image for DepX: AI-Powered DevOps Copilot
DepX: AI-Powered DevOps Copilot
2
81
Cover image for Revolutionize Your Sales Process with Instant Deal Closures
you have been WARNED. I am going to make every contact form, every Calendly link, and every "we'll get back to you" completely irrelevant. Conversations close deals. I made them instant. converc.com (http://converc.com)
0
20
Cover image for DailyCue — AI-Powered Decision Support
DailyCue — AI-Powered Decision Support for Recruitment Teams URL: https://dailycue.tech One-liner Shipped a recruitment SaaS MVP in 3 days — AI that tells recruiters exactly who to call today, and why. Project Overview Most recruitment teams don't have a lead problem. They have a prioritisation problem. CRMs are full, but consultants still rely on gut feel to decide where to focus each day — which means high-value opportunities get buried and revenue stays unpredictable. DailyCue connects to a recruitment CRM and uses AI to surface daily, prioritised actions for each consultant: who to contact, what the context is, and why now. It replaces the daily morning scramble with a clear, data-driven cue. What I Built CRM data ingestion pipeline with normalisation and deduplication AI prompt layer (GPT-4) generating personalised daily action lists per consultant Tiered Stripe billing in GBP with seat-based pricing JobAdder marketplace OAuth integration for one-click CRM connection Serverless architecture on Vercel with Inngest for background job processing Landing page, onboarding flow, and trial signup How I Shipped It Initial version went live in 3 days. I then iterated directly with real recruitment users, refining the AI prompts based on actual usage feedback until the output matched how experienced consultants naturally think about their pipeline. The product continued to evolve through multiple rounds of user testing and positioning pivots. Tech Stack Next.js · Supabase · PostgreSQL · OpenAI API · Stripe · Inngest · Vercel · JobAdder API
1
2
185
Cover image for Converc — Turn Website Visitors
Converc — Turn Website Visitors Into Live Conversations URL: https://converc.com One-liner Built a real-time browser-based calling platform from scratch — MVP live in 5 days. Project Overview Converc solves a problem that costs B2B companies revenue every day: high-intent visitors land on your site, can't get a human immediately, and leave. It embeds a call widget that connects visitors to sales reps in real time — no booking links, no forms, no friction. I architected and built the full product: a Next.js frontend, Supabase backend with Row Level Security, and WebRTC-powered peer-to-peer audio calling that works entirely in the browser with no native app required. The stack was chosen for performance, low latency, and the ability to ship fast without sacrificing production quality. What I Built WebRTC call engine with real-time signalling via Supabase Realtime Embeddable call widget (drop-in script for any website) Agent dashboard with live call status, queue management, and session history Webhook integration layer for CRM and Slack notifications on call events OAuth-based authentication with Google and dev/prod environment separation Supabase RLS policies enforcing strict data isolation between workspaces How I Shipped It The core MVP — widget, signalling, agent dashboard, and working calls — was shipped in 5 days using Cursor with Claude as the AI pair programmer. Post-MVP work covered Google Safe Browsing clearance, Slack App submission, analytics scoping, and hardening the auth and security model for production. Tech Stack Next.js · Supabase · PostgreSQL · WebRTC · Vercel · Tailwind CSS
0
148
LangChain(1)