You will get Generative AI assistant with custom language support (Telugu, Tenglish)
Project details
You will get a custom generative AI assistant that answers questions
in Telugu, English, or Tenglish over your own data (Excel, CSV, or
SQLite). I have built and deployed three live trilingual assistants
for government water resources, health schemes, and agriculture —
proven on real datasets: 116 reservoirs, 38,628 tanks, 6,214 medical
procedures, and 5,095 hospitals. I use Anthropic Claude APIs,
Text-to-SQL for structured data, and FastAPI/Streamlit for deployment.
Optional WhatsApp integration via Twilio is available. Source code and
deployment instructions included from Standard tier onward.
in Telugu, English, or Tenglish over your own data (Excel, CSV, or
SQLite). I have built and deployed three live trilingual assistants
for government water resources, health schemes, and agriculture —
proven on real datasets: 116 reservoirs, 38,628 tanks, 6,214 medical
procedures, and 5,095 hospitals. I use Anthropic Claude APIs,
Text-to-SQL for structured data, and FastAPI/Streamlit for deployment.
Optional WhatsApp integration via Twilio is available. Source code and
deployment instructions included from Standard tier onward.
AI Algorithms
Large Language Model, Transformer ModelAI Applications
AI Chatbot, Conversational AI, Machine Translation, Natural Language Generation, Natural Language UnderstandingAI Development Language
PythonAI Tools
Hugging Face, StreamlitWhat's included
| Service Tiers |
Starter
$600
|
Standard
$1,500
|
Advanced
$2,800
|
|---|---|---|---|
| Delivery Time | 5 days | 10 days | 14 days |
Number of Revisions | 1 | 3 | 4 |
AI Model Integration | - | - | - |
Batch Normalization | - | - | - |
Database Integration | |||
Detailed Code Comments | - | ||
Image Upscaling | - | - | - |
MLOps | - | - | - |
Model Deployment | - | ||
Model Documentation | - | - | - |
Model Monitoring | - | - | - |
Model Testing & Optimization | - | - | |
Model Tuning | - | - | - |
Natural Language Processing | |||
NLP Tokenization | - | - | - |
Pre-Training | - | - | - |
Prompt Engineering | |||
Setup File | - | - | - |
Source Code | - |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$150 - $400
WhatsApp Integration
(+ 2 Days)
+$400Frequently asked questions
About Abhishek
AI Agent Dev | 3 Live GovTech Systems | Claude API & Multilingual NLP
Visakhapatnam, India - 5:29 am local time
I build the ones that don't.
→ JalSathi: Trilingual water resource assistant — 116 reservoirs, 265 river gauges, 38,628 MI
tanks, 28-district rainfall. Dual-model Claude routing (Haiku for intent, Opus for
Telugu/Tenglish output). Production-deployed on Render.
→ VaidyaMitra: Government health scheme navigator — 6,214 procedures, 5,095 hospitals,
WhatsApp delivery via Twilio. Custom SQL post-processing for scheme-tier pricing edge cases.
→ KisanMitra: Crop + weather guidance in Telugu/English. 1,300+ impressions at public launch.
Data engineering backbone:
→ AWS serverless pipelines: 97.9–98.5% data quality, validated at row level
→ Kafka + Flink SQL real-time streaming on Confluent Cloud
→ Databricks + Spark: 56,934-row Delta Lake transformations, Unity Catalog
→ Airflow, Snowflake, Docker — orchestrated end-to-end
98% — Codebasics Data Engineering Bootcamp | 92% — AtliQ Virtual Internship | NIT Warangal
We're a strong fit if you're:
✅ Building AI for Indian government, civic tech, or regional language users
✅ Integrating Claude API with routing logic and production deployment
✅ Running pipelines where DQ, schema drift, and monitoring actually matter
✅ In a domain where your developer needs to understand the data — not just move it
Not the right fit if:
✗ You need a no-code chatbot by tomorrow
✗ The scope doesn't justify production-grade engineering
✗ You want someone who won't push back when shortcuts compromise reliability
Working with me, you get:
★ A developer who reads the domain data before touching the API
★ Production systems with real error handling, dual-model routing, and schema validation
★ Telugu, Hindi, and English — for the system and for the user
Message me now — tell me what you're building and we'll figure out fit in one conversation.
Steps for completing your project
After purchasing the project, send requirements so Abhishek can start the project.
Delivery time starts when Abhishek receives requirements from you.
Abhishek works on your project following the steps below.
Revisions may occur after the delivery date.
Data preparation and Text-to-SQL setup
Clean and validate client data, infer schema, build SQLite database, and design Text-to-SQL agent. (As done with 21 APWRIMS Excel files in JalSathi.)
AI agent development and testing
Implement dual-model routing (Claude Haiku for intent, Claude Opus for response generation), test with client's example questions in target languages.