Harika Sri
I am a dedicated professional driven by a blend of logical rigor and creative problem-solving, always striving for mastery in my field. My ambitious pursuit of innovative solutions is underpinned by a diligent focus on practical outcomes, ensuring precision and impact in every endeavor.
Identity verified by DigiLocker
Contact details — on requestProof of Work — on request
Experience
Document Artifact Expert
AfterQuery Experts · Apr 2026 – May 2026
- Produced high-quality document artifacts (PPTX, DOCX, PDF, XLSX) for large-scale Al training data collection as part of a vetted expert network.
- Applied domain expertise in data science and evaluation standards to ensure artifact quality and instructional alignment for frontier model development.
Document Artifact
AfterQuery Experts · Apr 2026 – May 2026
- • Submitting high-quality document artifacts (PPTX, DOCX, PDF, XLSX) for large-scale AI training data
- collection.
- • Contributing as part of a vetted network of AI evaluation experts sourcing structured documents for frontier
- model development.
- • Applying domain expertise in data science, evaluation, and document standards to ensure artifact quality
- and instructional alignment.
Business Analyst (AI Evaluation)
Turing · Jan 2026 – Apr 2026
- • Executing Side-by-Side (SxS) evaluations for Browser-Based AI Agents, assessing Navigation,
- Interaction, and Efficiency.
- • Conducting rigorous Fact-Checking to validate information and identify hallucinations without
- reliance on LLMs.
- • Providing detailed, structured justifications for ratings (Completion, Factual Accuracy, AI
- Performance) to support model improvements.
- • Maintained 100% adherence to strict evaluation guidelines while utilizing browser-based tools and
- extensions.
Business Analyst - AI Evaluation
Turing · Jan 2026 – Apr 2026
- Executed Side-by-Side (SxS) evaluations for Browser-Based Al Agents across Navigation, Interaction, and Efficiency dimensions.
- Conducted rigorous fact-checking to identify hallucinations without reliance on LLMs; maintained 100% adherence to strict evaluation guidelines.
- Provided structured justifications for ratings across Completion, Factual Accuracy, and Al Performance to support model improvement pipelines.
Vibe Coder
Mercor · Feb 2026 – Mar 2026
- • Spearheading the Vibe Coding Project to train models in building client-side web applications
- (React/HTML/JS).
- • Designing high-quality Prompts and Atomic Rubrics (8-12 tags) to teach models real-world skills like
- state management, event handling, and layout correctness.
- • Calibrating task difficulty to ensure model pass rates align with intended complexity (Easy: 70–100%,
- Medium: 40–70%, Hard: 0–40%).
- • Generating metadata and validating Vision-Language Models using single-shot screenshots and code
- inspection.
STEM OCR Specialist
Mercor · Feb 2026 – Mar 2026
- • Validating and repairing math/STEM problem-solution pairs extracted via OCR from textbook PDFs,
- ensuring 100% alignment with source material.
- • Standardizing complex LaTeX formatting for symbols, subscripts, and multi-step equations to ensure
- accurate rendering for AI training datasets.
- • Reformulating demonstrative content and ”not self-contained” problems into structured QA pairs to
- improve model instruction adherence.
- • Assessing the logical consistency and completeness of multi-step mathematical reasoning, including
- proof-based exercises and Chain-of-Thought (CoT) solutions.
Vibe Coder STEM OCR Specialist Generalist AI Trainer
Mercor · Nov 2025 – Mar 2026
- Designed atomic rubrics (8-12 tags) with difficulty calibration (Easy 70–100% / Medium 40–70% / Hard 0-40%) to train models in building client-side web applications.
- Validated Vision-Language Models using single-shot screenshots and code inspection; generated metadata for React/HTML/JS tasks.
- Repaired OCR-extracted STEM problem-solution pairs, standardised LaTeX formatting, and assessed Chain-of-Thought (CoT) reasoning completeness.
- Authored Golden Responses and multi-turn conversations with structural and behavioural constraints (JSON, Persona, Tone) for RLHF fine-tuning.
- Conducted live evaluation of Al systems (Google AIM vs.
- ChatGPT) during Cricket matches testing hallucinations on Betting Markets, Fantasy Leagues, and Player Profiles.
Generalist
Mercor · Jan 2026 – Feb 2026
- • Developed complex System Instructions (SI) and multi-turn conversations to train Generalist AI
- agents.
- • Implemented Structural & Behavioral Constraints (e.g., JSON format, Persona, Tone) to test model
- adherence.
- • Authored ”Golden Responses” demonstrating perfect instruction adherence for RLHF fine-tuning.
- • Managed prompt creation across domains (Finance, Coding, Generalist) ensuring diversity in training
- data.
Cricket Domain Expert (Live Sports Evaluation)
Mercor · Dec 2025 – Jan 2026
- • Executed live, dynamic evaluation of AI systems (Google AIM vs.
- ChatGPT) by generating authentic, game-aware queries during live Cricket matches.
- • Engineered high-complexity prompts covering Betting Markets, Fantasy Leagues, Historical Stats, and Player Profiles to test model hallucinations in real-time.
- • Authored detailed, objective justifications (in professional 3rd-person tone) assessing model performance based on Factuality, Helpfulness, and Instruction Adherence.
- • Validated AI responses against live broadcast data to identify inconsistencies in score reporting and momentum-based analysis.
STEM Evaluator - Math (SME)
Mercor · Nov 2025 – Jan 2026
- • Acting as a STEM SME to evaluate high-complexity mathematical training data for LLM optimization.
- • Validating Chain of Thought (CoT) reasoning to ensure AI models generate logically sound, step-by-step solutions.
- • Conducting QA (Quality Assurance) on training datasets, identifying hallucinations, and correcting
- derivation errors.
Data Annotator & Content Moderator
Innodata Inc. · Nov 2025 – Dec 2025
- Executed SFT annotation tasks for ML pipelines; refined Al-generated text for fluency, bias elimination, and PII compliance.
Data Annotator & Content Moderator
Innodata Inc. · Nov 2025 – Dec 2025
- 1.
- Executing SFT (Supervised Fine-Tuning) tasks by annotating and labeling large-scale datasets for ML
- pipelines.
- 2.
- Refining AI-generated text for linguistic fluency, ensuring outputs are free from bias and adhere to PII safety protocols.
Assistant Statistical Officer
CEGIS Govt. of Telangana · Aug 2019 – Oct 2025
- Led digitisation of service records for ~10 lakh government employees across 29 departments and 199 Heads of Department - coordinating directly with the Finance Secretary.
- Architected Telangana's Digital Workforce Repository, reducing reporting time by 40% and enabling data-driven recruitment and vacancy planning.
- Applied ARIMA, Tax Buoyancy, Logarithmic Regression, and Consensus Forecasting to project Telangana's tax revenue Consensus Forecasting achieved sub-30 crore error on an ₹87,000 crore base.
- Conducted comparative interstate budget analysis using Time-Series methods; designed statistical dashboards for senior non-technical stakeholders.
Center for effective governance of Indian states (CEGIS)
Government of Telangana · Aug 2019 – Oct 2025
- • Statistical Modeling: Spearheaded the evaluation of state schemes and crop-cutting experiments using
- predictive modeling to optimize resource allocation.
- • Data Management: Architected Telangana’s Digital Repository, processing and cleaning large-scale
- datasets to improve data accessibility and reduce reporting time by 40%.
- • Strategic Analysis: Conducted comparative interstate analysis for budget assessments, utilizing Time-
- Series Analysis to identify fiscal trends and drive policy decisions.
- • Reporting: Designed statistical reports and visual dashboards to present complex data insights to non-
- technical stakeholders.
AI Data Evaluator (Hindi/English)
OneForma · Jun 2024 – Aug 2024
- Executed RLHF comparative ranking (SxS) evaluations of Al responses across accuracy and cultural nuance in Hindi and English.
Prompt Engineer
Soul AI · Mar 2024 – May 2024
- Engineered and refined prompts for RLHF training; reviewed outputs for technical accuracy, directly reducing model error rates.
Subject Matter Expert - Statistics
Chegg Inc. · May 2018 – Nov 2020
- Authored detailed step-by-step solutions for advanced Statistics and Probability problems; maintained high accuracy under strict academic formatting standards.
Skills 0 proven through work
Also works with
Analytical ThinkingStatistical InferenceStakeholder ManagementTeamworkOn-time DeliveryReceiving and Incorporating FeedbackAdaptabilityWork Planning and PrioritizationMachine Learning ApplicationData VisualizationGame Application DevelopmentProduct DevelopmentAI Agent DevelopmentWebsite UsageWeb ResearchAI Hallucination DetectionLLM Output Verification and ValidationModel Training ExecutionPrompt EngineeringCodingReinforcement LearningAI Model TestingContent CurationFollowing InstructionsSOP DevelopmentSoftware Quality AssuranceAnnotation Rubric DevelopmentInformation ResearchForecastingExploratory Data Analysis (EDA)Training Data CreationDocument FormattingTechnical Documentation
Proof of Work
Proof of Work
Harika shares this with people who ask. You'll hear back either way.
Education
M.Sc., Statistics
Indian Institute of Technology Kanpur (IIT Kanpur) · 2016 — 2018
B.Sc., Mathematics, Statistics & Computer Science
Osmania University · 2014 — 2016
Projects
Contact details
Contact details
Harika shares this with people who ask. You'll hear back either way.
Ask about this candidate
AI responses may contain errors. Proof status and source information are provided by Proof of Skill.
A Proof CV — one profile, kept current, with the proof attached. Back to the portal