Skip to main content
Challenge

Build a Multilingual Public Service Policy Navigator with Bhashini

Translate regional citizen service inquiries using simulated Bhashini translation pipelines to assess mock eligibility rules. Generate localized qualification explanations and step-by-step documentation checklists for community field facilitators.

Workflow AutomationHosted by Vera
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Translate regional citizen service inquiries using simulated Bhashini translation pipelines to assess mock eligibility rules. Generate localized qualification explanations and step-by-step documentation checklists for community field facilitators.

Gram panchayat service facilitator

Ingest citizen inquiry text in Indian languages and map ISO-639 language codes via simulated Bhashini pipeline configs.

Normalize inquiry criteria such as landholding size, annual household income, and resident state.

Evaluate parsed criteria against mock state public benefit policy rules without external government calls.

Generate a translated eligibility determination report with specific evidentiary document requirements.

Multi-language web form showing original input, Bhashini translation trace, policy rule outcomes, and checklist.

Create synthetic eligibility policies and synthetic citizen profiles in Hindi, Marathi, and Tamil with English reference translations. Starter examples below are synthetic. They are not downloaded records.

Bhashini documentation Utilizes Bhashini pipeline concepts (Pipeline Config, NMT tasks, ISO-639 language codes) to structure multilingual query flows.

dspy, pydantic, fastapi. DSPy systematically optimizes prompt pipelines to reliably extract structured policy parameters from informal vernacular text. Instruction-tuned multilingual model capable of processing Indic text and generating strict JSON outputs.

Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.

Correctly identifies ISO-639 source language and maps to mock Bhashini translation pipeline.

Accurately extracts numeric income and landholding variables into structured Pydantic models.

Emits clear eligibility outcome and missing prerequisite documents in both English and target language.

```json [ { "input": { "source_lang": "hi", "query_text": "मेरे पास 1.5 एकड़ ज़मीन है और वार्षिक आय 120000 रुपये है, क्या मैं सोलर पंप योजना के पात्र हूँ?", "policy_id": "solar_irrigation_2025" }, "expected_output": { "target_lang": "hi", "normalized_criteria": { "landholding_acres": 1.5, "annual_income_inr": 120000 }, "eligibility_status": "ELIGIBLE", "reasoning": "Landholding is under the 5-acre limit and annual income is below 200,000 INR limit.", "required_documents": [ "Land Ownership Record (Khasra/Khatauni)", "Income Certificate", "Electricity Connection NOC" ] } }, { "input": { "source_lang": "mr", "query_text": "माझे वार्षिक उत्पन्न 450000 रुपये आहे आणि 6 एकर शेती आहे.", "policy_id": "solar_irrigation_2025" }, "expected_output": { "target_lang": "mr", "normalized_criteria": { "landholding_acres": 6, "annual_income_inr": 450000 }, "eligibility_status": "INELIGIBLE", "reasoning": "Landholding exceeds maximum limit of 5 acres and income exceeds 200,000 INR limit.", "required_documents": [] } } ] ```

Mock benefit policies only; no official Aadhaar verification or real administrative determinations are performed. Use mock external writes. Require human review before any real deployment.

Rules

Participation guide

Mock benefit policies only; no official Aadhaar verification or real administrative determinations are performed. Use mock external writes and human review.

Delivery guide

How work is evaluated

Evaluation

Correctly identifies ISO-639 source language and maps to mock Bhashini translation pipeline. Accurately extracts numeric income and landholding variables into structured Pydantic models. Emits clear eligibility outcome and missing prerequisite documents in both English and target language.

Submission format

Runnable source, pinned dependencies, setup commands, working interface, example outputs, recorded demo, tests, and action logs.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
4 scoring checks
Binary
1 pass or fail dimensions
Ordinal
3 scaled dimensions
Dimension 1

Complete workflow

Ingest citizen inquiry text in Indian languages and map ISO-639 language codes via simulated Bhashini pipeline configs. Normalize inquiry criteria such as landholding size, annual household income, and resident state. Evaluate parsed criteria against mock state public benefit policy rules without external government calls. Generate a translated eligibility determination report with specific evidentiary document requirements.

Score scale
012
Score 0

Missing or incorrect.

Score 1

Partly complete, with documented gaps.

Score 2

Complete, with reproducible evidence.

Dimension 2

Correctness and evidence

Correctly identifies ISO-639 source language and maps to mock Bhashini translation pipeline. Accurately extracts numeric income and landholding variables into structured Pydantic models. Emits clear eligibility outcome and missing prerequisite documents in both English and target language.

Score scale
012
Score 0

Missing or incorrect.

Score 1

Partly complete, with documented gaps.

Score 2

Complete, with reproducible evidence.

Dimension 3

Review interface

Multi-language web form showing original input, Bhashini translation trace, policy rule outcomes, and checklist.

Score scale
012
Score 0

Missing or incorrect.

Score 1

Partly complete, with documented gaps.

Score 2

Complete, with reproducible evidence.

Dimension 4

Safe execution

Mock benefit policies only; no official Aadhaar verification or real administrative determinations are performed.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Ingest citizen inquiry text in Indian languages and map ISO-639 language codes via simulated Bhashini pipeline configs.

  • Normalize inquiry criteria such as landholding size, annual household income, and resident state.

  • Evaluate parsed criteria against mock state public benefit policy rules without external government calls.

  • Generate a translated eligibility determination report with specific evidentiary document requirements.

Resources and assets

Reference links and supporting material

Dataset notes

Create synthetic eligibility policies and synthetic citizen profiles in Hindi, Marathi, and Tamil with English reference translations. Supplied starter examples are synthetic, not source extracts.

Challenge setup

Configuration summary for this challenge.

Challenge input

Input details are not configured.

DSPy

Prompt and LM program optimization framework.

Evaluated output

No output checks are configured.

Evaluation checks
  • Ingest citizen inquiry text in Indian languages and map ISO-639 language codes via simulated Bhashini pipel...
  • Correctly identifies ISO-639 source language and maps to mock Bhashini translation pipeline. Accurately ext...
  • Multi-language web form showing original input, Bhashini translation trace, policy rule outcomes, and check...
Configured evaluation evidence
  • 2 public reference cases
Configured runtime
  • 3 configured tools

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Action Space
  • DSPy · Optional
Training Infra
  • DSPy · Optional
Orchestration
  • DSPy · Optional

Evaluation contract

  • Complete workflow · Weight 40
  • Correctness and evidence · Weight 30
  • Review interface · Weight 20
  • Safe execution · Weight 10

Reference cases

  • 2 public reference cases

Workflow tasks

No workflow tasks are configured.

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a Multilingual Public Service Policy Navigator with Bhashini